div.ProseMirror

GraphDTA 실습 예제

1. Library

1-1. Import

본 예제에서 사용하는 모든 라이브러리, 패키지 및 모듈 import 와 RDKit 사용을 위한 시스템 경로 설정을 작성합니다.사용하는 라이브러리, 패키지 및 모듈에 대한 설명은 다음과 같습니다.

sys : 파이썬 인터프리터가 제공하는 변수와 함수를 직접 제어할 수 있게 해주는 모듈입니다.

os : 환경 변수나 디렉터리, 파일 등의 OS 자원을 제어할 수 있게 해주는 모듈입니다.

numpy : 벡터 및 행렬 연산에 있어서 매우 편리한 기능을 제공하는 라이브러리입니다.

pandas : 파이썬 데이터 처리를 위한 라이브러리로 numpy와 함께 데이터 분석에 있어 필수 라이브러리로 사용되고 있습니다.

json : python 타입을 json 형태의 문자열로 바꾸거나 반대의 기능을 제공하는 모듈입니다.

pickle : 객체의 형태를 그대로 유지하면서 파일에 저장하고 불러올 수 있게 하는 모듈입니다.

networkx : 그래프를 다루기 위한 라이브러리입니다.

math : 복잡한 연산을 다루기 위한 모듈입니다.

random : 임의의 숫자를 생성하거나 다양한 랜덤 관련 함수를 제공하는 모듈입니다.

collections : 튜플, 딕셔너리 객체에 대한 확장 데이터 구조를 제공하는 모듈입니다.

scipy : 과학 기술 계산용 함수 및 알고리즘을 제공하는 라이브러리입니다.

torch : 파이썬을 기반으로 하는 Scientific Computing 패키지이며 GPU를 제대로 이용하기 위한 numpy의 대체제로 사용하거나 딥 러닝 연구 플랫폼으로 사용합니다.

torch_geometric : 그래프 데이터 핸들링 및 학습에 사용하는 라이브러리입니다. 해당 라이브러리를 사용하기 위해선 pytorch와 cuda의 버전을 맞춰줘야 하며, 본 예제에서 사용하는 pytorch 버전은 1.6.0, cuda 버전은 10.2으로 설정했습니다.

rdkit : 화학물질의 정보를 담고 있는 파일형식의 데이터를 이용해서 화학물질의 구조이미지 구조 식을 만들어내는 패키지입니다.

import sys
sys.path.append('/usr/local/lib/python3.7/site-packages/')
sys.path.append('/opt/conda/lib/python3.7/site-packages/')
import os
import numpy as np
import pandas as pd
import json,pickle
import networkx as nx
from math import sqrt
from random import shuffle
from collections import OrderedDict
from scipy import stats
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.nn import Sequential, Linear, ReLU
from rdkit import Chem
from rdkit.Chem import MolFromSmiles
from torch_geometric import data as DATA
from torch_geometric.data import InMemoryDataset, DataLoader
from torch_geometric.nn import GCNConv, global_max_pool as gmp
from torch_geometric.nn import GCNConv, GATConv, GINConv, global_add_pool
from torch_geometric.nn import global_mean_pool as gap, global_max_pool as gmp
# 시각화 라이브러리
from matplotlib import pyplot as plt
import seaborn as sns
%matplotlib inline
7.9s
Python
→

데이터 및 사전 학습 파일을 불러오기 위해 wget 명령어로 파일을 다운로드 받습니다.

wget bit.ly/3hadHIb
15.6s
Bash in Python
→

/content/drive/My Drive/GraphDTA 라는 디렉토리를 생성하고, 다운로드받은 파일을 해당 경로에 압축해제합니다.

mkdir -p /content/drive/My\ Drive/GraphDTA
unzip 3hadHIb -d /content/drive/My\ Drive/GraphDTA
16.6s
Bash in Python
→

2. Create Data

2.1 Code

원자 특성 정의(one-hot encoding)one-hot encoding : 단 하나의 값만 True, 나머지는 모두 False로 치환하는 인코딩

One-hot encoding.png
def one_of_k_encoding(x, allowable_set):
    if x not in allowable_set:
        raise Exception("input {0} not in allowable set{1}:".format(x, allowable_set))
    return list(map(lambda s: x == s, allowable_set))
0.0s
Python
→
def one_of_k_encoding_unk(x, allowable_set):
    """Maps inputs not in the allowable set to the last element."""
    if x not in allowable_set:
        x = allowable_set[-1]
    return list(map(lambda s: x == s, allowable_set))
0.0s
Python
→

Symbol(44), Degree(11), TotalNumHs(11), ImplicitValence(11), IsAromatic(1) 정보를 이어붙여 총 78차원의 원자 특성 생성

def atom_features(atom):
    return np.array(one_of_k_encoding_unk(atom.GetSymbol(),['C', 'N', 'O', 'S', 'F', 'Si', 'P', 'Cl', 'Br', 'Mg', 'Na','Ca', 'Fe', 'As', 'Al', 'I', 'B', 'V', 'K', 'Tl', 'Yb','Sb', 'Sn', 'Ag', 'Pd', 'Co', 'Se', 'Ti', 'Zn', 'H','Li', 'Ge', 'Cu', 'Au', 'Ni', 'Cd', 'In', 'Mn', 'Zr','Cr', 'Pt', 'Hg', 'Pb', 'Unknown']) +
                    one_of_k_encoding(atom.GetDegree(), [0, 1, 2, 3, 4, 5, 6,7,8,9,10]) +
                    one_of_k_encoding_unk(atom.GetTotalNumHs(), [0, 1, 2, 3, 4, 5, 6,7,8,9,10]) +
                    one_of_k_encoding_unk(atom.GetImplicitValence(), [0, 1, 2, 3, 4, 5, 6,7,8,9,10]) +
                    [atom.GetIsAromatic()])
0.0s
Python
→

one-hot encoding과 atom_features method에 대한 예제

SMILES 문자열 하나를 rdkit.Chem.MolFromSmiles를 이용해 분자 그래프 형식으로 변환합니다. 

example_smile = 'O=C(NC1CCNCC1)c1[nH]ncc1NC(=O)c1c(Cl)cccc1Cl'
example_mol = Chem.MolFromSmiles(example_smile)
0.0s
Python
→

예시 분자는 25개의 원자로 구성되어 있음을 알 수 있습니다.

len(example_mol.GetAtoms())
0.0s
Python
→
25

25개의 원자는 0부터 24까지의 index가 부여가 되는데요, 따라서 example_atom_index 변수의 값을 0부터 24 사이의 임의의 값으로 바꾸어 예제를 실행해보도록 하겠습니다.

example_atom_index = 0
0.0s
Python
→
example_feature = atom_features(example_mol.GetAtoms()[example_atom_index])
example_feature
0.0s
Python
→
array([False, False, True, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, True, False, False, False, False, False, False, False, False, False, True, False, False, False, False, False, False, False, False, False, False, True, False, False, False, False, False, False, False, False, False, False, False])
# Node feature - symbol
example_feature[:44]
0.0s
Python
→
array([False, False, True, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False, False])
# Node feature - degree
example_feature[44:55]
0.0s
Python
→
array([False, True, False, False, False, False, False, False, False, False, False])
# Node feature - total_num_Hs
example_feature[55:66]
0.0s
Python
→
array([ True, False, False, False, False, False, False, False, False, False, False])
# Node feature - implicit_num_Hs
example_feature[66:77]
0.0s
Python
→
array([ True, False, False, False, False, False, False, False, False, False, False])
# Node feature - is_aromatic
example_feature[77:]
0.0s
Python
→
array([False])

SMILES 문자열로부터 그래프 데이터(행렬) 생성

# Returns : 원자 개수, 원자 특성 행렬, 인접 행렬
def smile_to_graph(smile):
    # SMILES 문자열로부터 분자 그래프 데이터 생성
    mol = Chem.MolFromSmiles(smile)
    
    # mol.GetNumAtoms() : 분자에 소속되어 있는 원자의 개수
    c_size = mol.GetNumAtoms()
    
    features = []
    # 분자에 소속되어 있는 원자들을 순회하면서 원자 특성 정보 수집
    for atom in mol.GetAtoms():
        feature = atom_features(atom)
        # edge nomalization
        features.append( feature / sum(feature) )
    edges = []
    # 분자를 이루는 원자들의 연결 구조 정보를 순회하면서 인접 정보 수집
    # 연결 구조 정보 : 시작 원자 index, 끝 원자 index
    for bond in mol.GetBonds():
        edges.append([bond.GetBeginAtomIdx(), bond.GetEndAtomIdx()])
    # 연결 구조 정보를 통한 방향 그래프 생성    
    g = nx.Graph(edges).to_directed()
    edge_index = []
    for e1, e2 in g.edges:
        edge_index.append([e1, e2])
  
    return c_size, features, edge_index
0.0s
Python
→

smile_to_graph method에 대한 예제

example_mol = Chem.MolFromSmiles(example_smile)
0.0s
Python
→

예시 분자를 구성하는 원자의 개수를 출력합니다.

example_c_size = example_mol.GetNumAtoms()
example_c_size
0.0s
Python
→
25

예시 분자를 구성하는 원자들의 특성 행렬을 출력합니다.

example_features = []
for atom in example_mol.GetAtoms():
    feature = atom_features(atom)
    example_features.append( feature / sum(feature) )
    
example_features
0.1s
Python
→

예시 분자들을 구성하는 원자들의 결합 정보를 출력합니다.

example_edges = []
for bond in example_mol.GetBonds():
    example_edges.append([bond.GetBeginAtomIdx(), bond.GetEndAtomIdx()])
example_edges
0.0s
Python
→
[[0, 1], [1, 2], [2, 3], [3, 4], [4, 5], [5, 6], [6, 7], [7, 8], [1, 9], [9, 10], [10, 11], [11, 12], [12, 13], [13, 14], [14, 15], [15, 16], [15, 17], [17, 18], [18, 19], [18, 20], [20, 21], [21, 22], [22, 23], [23, 24], [8, 3], [13, 9], [23, 17]]
print(len(example_edges), len(example_edges[0]))
0.3s
Python
→

원자 결합 정보를 바탕으로 nextworkx 라이브러리를 활용하여 방향 그래프로 데이터를 변환합니다.

example_graph = nx.Graph(example_edges).to_directed()
example_graph
0.0s
Python
→
<networkx.classes.digraph.DiGraph at 0x7fdcad5ed550>
example_edge_index = []
for e1, e2 in example_graph.edges:
    example_edge_index.append([e1, e2])
    
example_edge_index
0.0s
Python
→
[[0, 1], [1, 0], [1, 2], [1, 9], [2, 1], [2, 3], [3, 2], [3, 4], [3, 8], [4, 3], [4, 5], [5, 4], [5, 6], [6, 5], [6, 7], [7, 6], [7, 8], [8, 7], [8, 3], [9, 1], [9, 10], [9, 13], [10, 9], [10, 11], [11, 10], [11, 12], [12, 11], [12, 13], [13, 12], [13, 14], [13, 9], [14, 13], [14, 15], [15, 14], [15, 16], [15, 17], [16, 15], [17, 15], [17, 18], [17, 23], [18, 17], [18, 19], [18, 20], [19, 18], [20, 18], [20, 21], [21, 20], [21, 22], [22, 21], [22, 23], [23, 22], [23, 24], [23, 17], [24, 23]]
print(len(example_edge_index), len(example_edge_index[0]))
0.4s
Python
→

label encoding 예시입니다.

# 표적 염기서열을 이루는 알파벳(25자) vocabulary
seq_voc = "ABCDEFGHIKLMNOPQRSTUVWXYZ"
# 알파벳들을 정수(1 ~ 25)로 매핑
seq_dict = {v:(i+1) for i,v in enumerate(seq_voc)}
# 시퀀스 딕셔너리 길이 : 25
seq_dict_len = len(seq_dict)
# 패딩을 위한 시퀀스 최대 길이 정의
max_seq_len = 1000
0.0s
Python
→
seq_dict
0.0s
Python
→
{'A': 1, 'B': 2, 'C': 3, 'D': 4, 'E': 5, 'F': 6, 'G': 7, 'H': 8, 'I': 9, 'K': 10, 'L': 11, 'M': 12, 'N': 13, 'O': 14, 'P': 15, 'Q': 16, 'R': 17, 'S': 18, 'T': 19, 'U': 20, 'V': 21, 'W': 22, 'X': 23, 'Y': 24, 'Z': 25}
# Returns : 길이가 1000인 Protein Representation(Integer/label encoding)
def seq_cat(prot):
    #  크기가 1000 이고 원소를 0으로 채운 배열 생성
    x = np.zeros(max_seq_len)
    # 시퀀스 딕셔너리를 참조하여 단백질 서열 정보 입력
    for i, ch in enumerate(prot[:max_seq_len]): 
        x[i] = seq_dict[ch]
        
    return x
0.0s
Python
→

단백질 시퀀스 표현 예시입니다.

example_csv = pd.read_csv('/content/drive/My Drive/GraphDTA/data/davis_train.csv')
example_prot_seq = example_csv.target_sequence[0]
example_prot_seq
0.2s
Python
→
len(example_prot_seq)
0.0s
Python
→
1338
example_seq_cat = seq_cat(example_prot_seq)
example_seq_cat
0.0s
Python
→
len(example_seq_cat)
0.0s
Python
→
1000
all_prots = []
datasets = ['kiba','davis']
0.0s
Python
→
for dataset in datasets:
    print('convert data from DeepDTA for ', dataset)
    # 파일 경로 지정 및 로드
    fpath = '/content/drive/My Drive/GraphDTA/data/' + dataset + '/'
    train_fold = json.load(open(fpath + "folds/train_fold_setting1.txt"))
    train_fold = [ee for e in train_fold for ee in e ]
    valid_fold = json.load(open(fpath + "folds/test_fold_setting1.txt"))
    ligands = json.load(open(fpath + "ligands_can.txt"), object_pairs_hook=OrderedDict)
    proteins = json.load(open(fpath + "proteins.txt"), object_pairs_hook=OrderedDict)
    affinity = pickle.load(open(fpath + "Y","rb"), encoding='latin1')
    drugs = []
    prots = []
    for d in ligands.keys():
        # 입체 구조 정보를 포함한 SMILES 문자열
        lg = Chem.MolToSmiles(Chem.MolFromSmiles(ligands[d]),isomericSmiles=True)
        drugs.append(lg)
    for t in proteins.keys():
        # 단백질 서열 정보
        prots.append(proteins[t])
    # davis 데이터셋의 데이터일 경우, 결합 친화도 수치 조정
    if dataset == 'davis':
        affinity = [-np.log10(y/1e9) for y in affinity]
    affinity = np.asarray(affinity)
    opts = ['train','test']
    # 위에서 정의한 drugs와 prots 배열 내의 약물 및 표적 정보에 근거하여 학습 및 테스트 데이터 파일 생성
    for opt in opts:
        rows, cols = np.where(np.isnan(affinity)==False)  
        if opt=='train':
            rows,cols = rows[train_fold], cols[train_fold]
        elif opt=='test':
            rows,cols = rows[valid_fold], cols[valid_fold]
        with open('/content/drive/My Drive/GraphDTA/data/' + dataset + '_' + opt + '.csv', 'w') as f:
            f.write('compound_iso_smiles,target_sequence,affinity\n')
            for pair_ind in range(len(rows)):
                ls = []
                ls += [ drugs[rows[pair_ind]]  ]
                ls += [ prots[cols[pair_ind]]  ]
                ls += [ affinity[rows[pair_ind],cols[pair_ind]]  ]
                f.write(','.join(map(str,ls)) + '\n') 
    print('\ndataset:', dataset)
    print('train_fold:', len(train_fold))
    print('test_fold:', len(valid_fold))
    print('len(set(drugs)),len(set(prots)):', len(set(drugs)),len(set(prots)))
    all_prots += list(set(prots))
2.1s
Python
→
davis_train = pd.read_csv('/content/drive/My Drive/GraphDTA/data/davis_train.csv')
davis_test = pd.read_csv('/content/drive/My Drive/GraphDTA/data/davis_test.csv')
kiba_train = pd.read_csv('/content/drive/My Drive/GraphDTA/data/kiba_train.csv')
kiba_test = pd.read_csv('/content/drive/My Drive/GraphDTA/data/kiba_test.csv')
0.6s
Python
→
davis_train.head(5)
0.4s
Python
→
davis_test.head(5)
0.3s
Python
→
kiba_train.head(5)
0.4s
Python
→
kiba_train.head(5)
0.0s
Python
→
compound_iso_smiles = []
for dt_name in ['kiba','davis']:
    opts = ['train','test']
    for opt in opts:
        df = pd.read_csv('/content/drive/My Drive/GraphDTA/data/' + dt_name + '_' + opt + '.csv')
        compound_iso_smiles += list( df['compound_iso_smiles'] )
# 분자 정보 중복 제거
compound_iso_smiles = set(compound_iso_smiles)
smile_graph = {}
# SMILES 문자열을 그래프 형태로 변환하여 딕셔너리 형태로 저장
for smile in compound_iso_smiles:
    g = smile_to_graph(smile)
    smile_graph[smile] = g
18.0s
Python
→
class TestbedDataset(InMemoryDataset):
    def __init__(self, root='/tmp', dataset='davis', 
                 xd=None, xt=None, y=None, transform=None,
                 pre_transform=None,smile_graph=None):
        # 전처리된 데이터 저장을 위한 경로, 기본값은 '/tmp'
        super(TestbedDataset, self).__init__(root, transform, pre_transform)
        # 벤치마크 데이터 셋으로 기본값은 'davis'
        self.dataset = dataset
        # 전처리 데이터가 존재하면 프로세스를 건너 뛰고 데이터 로드
        if os.path.isfile(self.processed_paths[0]):
            print('Pre-processed data found: {}, loading ...'.format(self.processed_paths[0]))
            self.data, self.slices = torch.load(self.processed_paths[0])
        # 그렇지 않으면 프로세스를 실행하고 데이터 로드
        else:
            print('Pre-processed data {} not found, doing pre-processing...'.format(self.processed_paths[0]))
            self.process(xd, xt, y,smile_graph)
            self.data, self.slices = torch.load(self.processed_paths[0])
    @property
    def raw_file_names(self):
        pass
    @property
    def processed_file_names(self):
        return [self.dataset + '.pt']
    def download(self):
        pass
    def _download(self):
        pass
    def _process(self):
        if not os.path.exists(self.processed_dir):
            os.makedirs(self.processed_dir)
    # XD : SMILES 목록
    # XT : 인코딩된 단백질(표적) 목록(label encoding)
    # Y : 레이블 목록(결합 친화도)
    # Return : PyTorch-Geometric 형식으로 처리된 데이터
    def process(self, xd, xt, y,smile_graph):
        assert (len(xd) == len(xt) and len(xt) == len(y)), "The three lists must be the same length!"
        data_list = []
        data_len = len(xd)
        # Graph Convolutional Networks Model의 입력 형식에 맞게 데이터 변환
        # 행 단위로 반복(xd, xt, y)
        for i in range(data_len):
            print('Converting SMILES to graph: {}/{}'.format(i+1, data_len))
            smiles = xd[i]
            target = xt[i]
            labels = y[i]
            # smile_graph : RDKit을 활용하여 분자 그래프 표현으로 변환된 SMILES 값 
            c_size, features, edge_index = smile_graph[smiles]
            # GCNData : PyTorch Geometrics GCN 알고리즘에 대한 그래프 데이터
            # torch.Tensor : 텐서 자료형으로 Numpy의 배열(ndarray)와 유사한 자료형
            # x : 특성 행렬
            # edge_index : 인접 행렬
            # labels : 결합 친화도
            # target : 표적(단백질)
            GCNData = DATA.Data(x=torch.Tensor(features),
                                edge_index=torch.LongTensor(edge_index).transpose(1, 0),
                                y=torch.FloatTensor([labels]))
            GCNData.target = torch.LongTensor([target])
            # c_size : 하나의 compound에 소속되어 있는 원자의 개수
            GCNData.__setitem__('c_size', torch.LongTensor([c_size]))
            
            # 그래프, 레이블, 표적 서열을 데이터 목록에 추가
            data_list.append(GCNData)
        if self.pre_filter is not None:
            data_list = [data for data in data_list if self.pre_filter(data)]
        if self.pre_transform is not None:
            data_list = [self.pre_transform(data) for data in data_list]
        print('Graph construction done. Saving to file.')
        data, slices = self.collate(data_list)
        # 전처리 데이터 저장
        torch.save((data, slices), self.processed_paths[0])
0.0s
Python
→
datasets = ['davis','kiba']
# PyTorch 데이터 포맷으로 변환
for dataset in datasets:
    # 전처리 데이터 로드
    processed_data_file_train = '/content/drive/My Drive/GraphDTA/data/processed/' + dataset + '_train.pt'
    processed_data_file_test = '/content/drive/My Drive/GraphDTA/data/processed/' + dataset + '_test.pt'
    
    if ((not os.path.isfile(processed_data_file_train)) or (not os.path.isfile(processed_data_file_test))):
        df = pd.read_csv('/content/drive/My Drive/GraphDTA/data/' + dataset + '_train.csv')
        train_drugs, train_prots,  train_Y = list(df['compound_iso_smiles']),list(df['target_sequence']),list(df['affinity'])
        XT = [seq_cat(t) for t in train_prots]
        train_drugs, train_prots,  train_Y = np.asarray(train_drugs), np.asarray(XT), np.asarray(train_Y)
        df = pd.read_csv('/content/drive/My Drive/GraphDTA/data/' + dataset + '_test.csv')
        test_drugs, test_prots,  test_Y = list(df['compound_iso_smiles']),list(df['target_sequence']),list(df['affinity'])
        XT = [seq_cat(t) for t in test_prots]
        test_drugs, test_prots,  test_Y = np.asarray(test_drugs), np.asarray(XT), np.asarray(test_Y)
        # PyTorch Geometric 데이터 생성
        print('preparing ', dataset + '_train.pt in pytorch format!')
        train_data = TestbedDataset(root='/content/drive/My Drive/GraphDTA/data', dataset=dataset+'_train', xd=train_drugs, xt=train_prots, y=train_Y,smile_graph=smile_graph)
        print('preparing ', dataset + '_test.pt in pytorch format!')
        test_data = TestbedDataset(root='/content/drive/My Drive/GraphDTA/data', dataset=dataset+'_test', xd=test_drugs, xt=test_prots, y=test_Y,smile_graph=smile_graph)
        print(processed_data_file_train, ' and ', processed_data_file_test, ' have been created')        
    else:
        print(processed_data_file_train, ' and ', processed_data_file_test, ' are already created')
0.3s
Python
→

2.2 Data Exploration

davis_train.csv 파일을 예시로 데이터가 어떤 형식으로 모델의 입력으로 들어가는지 살펴보겠습니다.아래의 코드는 davis_train.csv 파일의 첫 번째 행을 예시로 코드를 작성했습니다.해당 파일의 다른 행을 보시려면 변수 row_index의 값을 0 ~ 25045 사이의 값으로 바꿔주시면 됩니다.

train_drugs, train_prots,  train_Y = list(davis_train['compound_iso_smiles']),list(davis_train['target_sequence']),list(davis_train['affinity'])
0.0s
Python
→
row_index = 0
0.0s
Python
→

SMILES 문자열 예시입니다.

train_drugs[row_index]
0.0s
Python
→
'O=C(NC1CCNCC1)c1[nH]ncc1NC(=O)c1c(Cl)cccc1Cl'

표적 시퀀스 예시입니다.

train_prots[row_index]
0.1s
Python
→

결합 친화도 예시입니다.

train_Y[row_index]
0.0s
Python
→
5.0

알파벳으로 구성된 표적 시퀀스를 정수형으로 변환합니다.

XT = [seq_cat(t) for t in train_prots]
3.7s
Python
→
XT[row_index]
0.1s
Python
→
train_drugs, train_prots,  train_Y = np.asarray(train_drugs), np.asarray(XT), np.asarray(train_Y)
0.1s
Python
→

SMILES 문자열을 원자 개수, 원자 특성 행렬, 인접 행렬로 이루어진 그래프 데이터 형식으로 변환합니다.

c_size, features, edge_index = smile_graph[train_drugs[row_index]]
0.0s
Python
→
c_size
0.0s
Python
→
25
features
0.1s
Python
→
print(len(features), len(features[0]))
0.4s
Python
→
edge_index
0.0s
Python
→
[[0, 1], [1, 0], [1, 2], [1, 9], [2, 1], [2, 3], [3, 2], [3, 4], [3, 8], [4, 3], [4, 5], [5, 4], [5, 6], [6, 5], [6, 7], [7, 6], [7, 8], [8, 7], [8, 3], [9, 1], [9, 10], [9, 13], [10, 9], [10, 11], [11, 10], [11, 12], [12, 11], [12, 13], [13, 12], [13, 14], [13, 9], [14, 13], [14, 15], [15, 14], [15, 16], [15, 17], [16, 15], [17, 15], [17, 18], [17, 23], [18, 17], [18, 19], [18, 20], [19, 18], [20, 18], [20, 21], [21, 20], [21, 22], [22, 21], [22, 23], [23, 22], [23, 24], [23, 17], [24, 23]]
len(edge_index)
0.0s
Python
→
54

원자 특성 데이터를 텐서 형식으로 저장합니다.

f = torch.Tensor(features)
f
0.0s
Python
→
tensor([[0.0000, 0.0000, 0.2500, ..., 0.0000, 0.0000, 0.0000], [0.2500, 0.0000, 0.0000, ..., 0.0000, 0.0000, 0.0000], [0.0000, 0.2500, 0.0000, ..., 0.0000, 0.0000, 0.0000], ..., [0.2000, 0.0000, 0.0000, ..., 0.0000, 0.0000, 0.2000], [0.2000, 0.0000, 0.0000, ..., 0.0000, 0.0000, 0.2000], [0.0000, 0.0000, 0.0000, ..., 0.0000, 0.0000, 0.0000]])
torch.Tensor(features).shape
0.0s
Python
→
torch.Size([25, 78])

인접 행렬을 전치 연산하여 텐서 형식으로 저장합니다.

e = torch.LongTensor(edge_index).transpose(1, 0)
e
0.0s
Python
→
tensor([[ 0, 1, 1, 1, 2, 2, 3, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13, 13, 14, 14, 15, 15, 15, 16, 17, 17, 17, 18, 18, 18, 19, 20, 20, 21, 21, 22, 22, 23, 23, 23, 24], [ 1, 0, 2, 9, 1, 3, 2, 4, 8, 3, 5, 4, 6, 5, 7, 6, 8, 7, 3, 1, 10, 13, 9, 11, 10, 12, 11, 13, 12, 14, 9, 13, 15, 14, 16, 17, 15, 15, 18, 23, 17, 19, 20, 18, 18, 21, 20, 22, 21, 23, 22, 24, 17, 23]])
torch.LongTensor(edge_index).transpose(1, 0).shape
0.0s
Python
→
torch.Size([2, 54])

결합 친화도 데이터를 텐서 형식으로 저장합니다.

l = torch.FloatTensor([train_Y[row_index]])
l
0.0s
Python
→
tensor([5.])
gd = DATA.Data(x=f, edge_index=e, y=l)
0.0s
Python
→
gd.target = torch.LongTensor([XT[row_index]])
gd.__setitem__('c_size', torch.LongTensor([c_size]))
0.0s
Python
→
gd
0.0s
Python
→
Data(c_size=[1], edge_index=[2, 54], target=[1, 1000], x=[25, 78], y=[1])

3. Performance Evaluation Index Definition

3-1. Mean Square Error

추정값의 평균 제곱 오차 또는 평균 제곱 편차는 오차의 제곱 평균, 즉 추정값과 실제 값 간의 평균 제곱 차를 측정

MSE.png
def mse(y,f):
    mse = ((y - f)**2).mean(axis=0)
    return mse
0.0s
Python
→

3-2. Root Mean Square Error

추정 값 또는 모델이 예측한 값과 실제 환경에서 관찰되는 값의 차이를 다룰 때 흔히 사용하는 측도

RMSE.png
def rmse(y,f):
    rmse = sqrt(((y - f)**2).mean(axis=0))
    return rmse
0.0s
Python
→

3-3. Pearson Correlation Coefficient

변수 X 와 Y 간의 선형 상관 관계를 계량화한 수치

pearson-formula.png
def pearson(y,f):
    rp = np.corrcoef(y, f)[0,1]
    return rp
0.0s
Python
→

3-4. Spearman's Rank Correlation Coefficient

상관 계수를 계산할 두 데이터의 실제 값 대신 순위 rank를 사용해 상관 계수를 계산하는 방식

Spearman.png
def spearman(y,f):
    rs = stats.spearmanr(y, f)[0]
    return rs
0.0s
Python
→

3-5. Concordance Index

샘플들을 실제 결합 친화도의 오름차순으로 나열하고, '각 샘플들보다 실제 결합 친화도가 높은 개수를 모두 더한 총합'과 '각 샘플들보다 결합 친화도가 높을 것으로 올바르게 예측된 샘플들의 개수를 모두 더한 총합'의 비율로 계산

cindex.png
def ci(y,f):
    ind = np.argsort(y)
    y = y[ind]
    f = f[ind]
    i = len(y)-1
    j = i-1
    z = 0.0
    S = 0.0
    while i > 0:
        while j >= 0:
            if y[i] > y[j]:
                z = z+1
                u = f[i] - f[j]
                if u > 0:
                    S = S + 1
                elif u == 0:
                    S = S + 0.5
            j = j - 1
        i = i - 1
        j = i-1
    ci = S/z
    return ci
0.0s
Python
→

4. Model

4-1. GCN

GraphDTA Model.png
# GCN based model
class GCNNet(torch.nn.Module):
    def __init__(self, n_output=1, n_filters=32, embed_dim=128,num_features_xd=78, num_features_xt=25, output_dim=128, dropout=0.2):
        super(GCNNet, self).__init__()
        # 약물 분자 표현을 위한 GCN 구성
        # 모델의 출력은 하나의 실수로 나타낸 결합 친화도 값입니다. 따라서 n_output=1 입니다.
        self.n_output = n_output
        # Graph Convolutional Networks
        
        # 첫 번째 인자는 in_channels로 각 입력 샘플의 크기를 나타냅니다.
        # N 개의 노드와 E개의 엣지로 구성된 그래프 graph(N,E)가 존재할 때, 각 노드를 d 차원으로 임베딩한다면, n * d 차원의 입력값을 구할 수 있습니다.
        # 여기서는 각 노드를 78 차원의 특성으로 나눴으므로 첫 번째 그래프 합성곱 층의 입력 샘플 크기는 78로 주어집니다.
        # 두 번째 인자는 out_channels로 각 출력 샘플의 크기를 나타냅니다.
        self.conv1 = GCNConv(num_features_xd, num_features_xd)
        # 다음 층으로 진행할수록 고차원의 특성을 뽑아내기 위해 출력 샘플의 크기를 두 배씩 늘려줍니다.
        self.conv2 = GCNConv(num_features_xd, num_features_xd*2)
        self.conv3 = GCNConv(num_features_xd*2, num_features_xd * 4)
        # Graph Convolution Layer를 통해 312개의 특성으로 출력된 값들을 입력으로 하여 Linear 함수를 통해 선형 변환을 거쳐 1024개의 출력으로 변환합니다(Fully-Connected).
        self.fc_g1 = torch.nn.Linear(num_features_xd*4, 1024)
        self.fc_g2 = torch.nn.Linear(1024, output_dim)
        # 활성화함수는 어떠한 신호를 입력받아 이를 적절한 처리를 하여 출력해주는 함수입니다.
        # 이를 통해 출력된 신호가 다음 단계에서 활성화 되는지를 결정합니다.
        # ReLU는 활성화함수의 일종으로, 0 미만의 값은 0으로, 0 이상의 값은 그대로 전달하는 함수입니다.
        self.relu = nn.ReLU()
        # Dropout은 딥러닝 시 과적합 문제를 해결하기 위해 은닉층 노드 중 일부를 제거하는 학습기법 입니다.
        # 은닉층 노드 중 일부를 제거함으로써 특정 뉴런의 가중치나 영향이 감소하므로 보다 강건한 신경망을 구성할 수 있게 해줍니다.
        self.dropout = nn.Dropout(dropout)
        # 단백질 서열 표현을 위한 1D Conv 층 구성
        # 학습 가능한 임베딩 테이블을 생성합니다.
        # 첫 번째 인자는 num_embeddings로 임베딩을 할 단어들의 개수, 즉 단어 집합의 크기입니다.
        # 1을 더해준 이유는 Create Data 부분에서 단백질 서열을 정수화 인코딩을 할 때 입력 크기를 맞추기 위한 패딩 0을 넣었습니다.
        # 따라서 1 ~ 25 의 정수로 이루어진 단어 집합에 0을 추가하여 총 26개의 단어 집합을 입력으로 합니다.
        # 두 번째 인자는 embedding_dim으로 임베딩을 할 벡터의 차원입니다. 해당 부분은 사용자가 정해주는 하이퍼 파라미터입니다.
        self.embedding_xt = nn.Embedding(num_features_xt + 1, embed_dim)
        # 각 단백질 서열은 길이 1000에 맞춰 패딩을 삽입하여 변환을 했습니다. 따라서 in_channels=1000 입니다.
        # out_channels는 각 출력 샘플의 크기를 나타냅니다. 여기서는 n_filter=32로 설정했습니다.
        # 단백질 서열을 구성하는 각 정수(단어)는 128 차원으로 임베딩되었습니다. 따라서 실제 출력되는 형식은 다음과 같습니다.
        # out_channels(32) * (embed_dim(128) - kernel_size(8) + 1) = 32*121
        self.conv_xt_1 = nn.Conv1d(in_channels=1000, out_channels=n_filters, kernel_size=8)
        # GCN Layer와 마찬가지로 32*121개의 특성으로 출력된 값들을 입력으로 하여 Linear 함수를 통해 선형 변환을 거쳐 output_dim개의 출력으로 변환합니다(Fully-Connected).
        self.fc1_xt = nn.Linear(32*121, output_dim)
        # GCN Layer를 통해 생성된 Drug Molecule Representation과 Conv1d Layer를 통해 생성된 Protein Representation을 결합하는 Fully-Connected Layer 입니다.
        # 두 Representation은 출력 크기가 output_dim으로 결합 크기는 output_dim*2가 됩니다.
        self.fc1 = nn.Linear(2*output_dim, 1024)
        self.fc2 = nn.Linear(1024, 512)
        # 최종으로 출력되는 값은 하나의 실수로 나타낸 결합 친화도 값입니다. 따라서 최종 출력 크기는 n_output=1 입니다.
        self.out = nn.Linear(512, self.n_output)
    def forward(self, data):
        # 약물 분자 입력
        # x : 원자 특성 행렬
        # edge_index : 인접 행렬
        # batch : batch size
        x, edge_index, batch = data.x, data.edge_index, data.batch
        # 단백질 서열 입력
        # target : 단백질 서열 정보
        target = data.target
        # Graph Convolutional Networks Layer
        x = self.conv1(x, edge_index)
        x = self.relu(x)
        x = self.conv2(x, edge_index)
        x = self.relu(x)
        x = self.conv3(x, edge_index)
        x = self.relu(x)
        # gmp : global max pooling
        x = gmp(x, batch)       
        # 활성화(self.relu), 전결합(self.fc_g1, self.fc_g2) 및 정규화(self.dropout)
        x = self.relu(self.fc_g1(x))
        x = self.dropout(x)
        x = self.fc_g2(x)
        x = self.dropout(x)
        # Convolutional Neural Networks Layer(1D)
        embedded_xt = self.embedding_xt(target)
        conv_xt = self.conv_xt_1(embedded_xt)
        # 행렬의 형식을 변환합니다.
        # 단백질 서열 정보가 입력되는 Conv1d Layer는 출력 형식이 1차원입니다.
        # 이를 완전 연결 층의 입력으로 사용하기 위해선 2차원으로 형식을 변환해주어야 합니다.
        # view는 텐서를 재구성해주는 함수이며, 아래 코드는 기존 1차원 형식의 출력값을 2차원으로 변형해주는 코드입니다.
        xt = conv_xt.view(-1, 32 * 121)
        xt = self.fc1_xt(xt)
        # GCN Layer를 통해 생성된 Drug Molecule Representation과 Conv1d Layer를 통해 생성된 Protein Representation을 torch.cat 함수를 통해 하나로 결합합니다.
        xc = torch.cat((x, xt), 1)
        # 최종으로 출력되는 값은 하나의 실수로 나타낸 결합 친화도 값입니다.
        # 따라서 전결합(fc1, fc2, out), 활성화(self.relu) 및 정규화(dropout)를 통해 하나의 값을 출력합니다.
        xc = self.fc1(xc)
        xc = self.relu(xc)
        xc = self.dropout(xc)
        xc = self.fc2(xc)
        xc = self.relu(xc)
        xc = self.dropout(xc)
        out = self.out(xc)
        return out
0.0s
Python
→

Activation Function(ReLU) 참고자료

ReLU.png

Dropout 참고자료

Dropout.png

4-2. GAT

# GAT based model
class GATNet(torch.nn.Module):
    def __init__(self, num_features_xd=78, n_output=1, num_features_xt=25,
                     n_filters=32, embed_dim=128, output_dim=128, dropout=0.2):
        super(GATNet, self).__init__()
        # 약물 분자 표현을 위한 GAT 층 구성
        self.gcn1 = GATConv(num_features_xd, num_features_xd, heads=10, dropout=dropout)
        self.gcn2 = GATConv(num_features_xd * 10, output_dim, dropout=dropout)
        self.fc_g1 = nn.Linear(output_dim, output_dim)
        # 단백질 서열 표현을 위한 1D Conv 층 구성
        self.embedding_xt = nn.Embedding(num_features_xt + 1, embed_dim)
        self.conv_xt1 = nn.Conv1d(in_channels=1000, out_channels=n_filters, kernel_size=8)
        self.fc_xt1 = nn.Linear(32*121, output_dim)
        # 층 결합
        self.fc1 = nn.Linear(256, 1024)
        self.fc2 = nn.Linear(1024, 256)
        self.out = nn.Linear(256, n_output)
        # 활성화 및 정규화(flatten)
        self.relu = nn.ReLU()
        self.dropout = nn.Dropout(dropout)
    def forward(self, data):
        x, edge_index, batch = data.x, data.edge_index, data.batch
        x = F.dropout(x, p=0.2, training=self.training)
        x = F.elu(self.gcn1(x, edge_index))
        x = F.dropout(x, p=0.2, training=self.training)
        x = self.gcn2(x, edge_index)
        x = self.relu(x)
        # gmp : global max pooling
        x = gmp(x, batch)          
        x = self.fc_g1(x)
        x = self.relu(x)
        target = data.target
        embedded_xt = self.embedding_xt(target)
        conv_xt = self.conv_xt1(embedded_xt)
        conv_xt = self.relu(conv_xt)
        # flatten
        xt = conv_xt.view(-1, 32 * 121)
        xt = self.fc_xt1(xt)
        # concat
        xc = torch.cat((x, xt), 1)
        # Dense 층 추가
        xc = self.fc1(xc)
        xc = self.relu(xc)
        xc = self.dropout(xc)
        xc = self.fc2(xc)
        xc = self.relu(xc)
        xc = self.dropout(xc)
        out = self.out(xc)
        return out
0.0s
Python
→

4-3. GAT_GCN

# GCN-CNN based model
class GAT_GCN(torch.nn.Module):
    def __init__(self, n_output=1, num_features_xd=78, num_features_xt=25,
                 n_filters=32, embed_dim=128, output_dim=128, dropout=0.2):
        super(GAT_GCN, self).__init__()
        # 약물 분자 표현을 위한 GAT 및 GCN 층 구성
        self.n_output = n_output
        self.conv1 = GATConv(num_features_xd, num_features_xd, heads=10)
        self.conv2 = GCNConv(num_features_xd*10, num_features_xd*10)
        self.fc_g1 = torch.nn.Linear(num_features_xd*10*2, 1500)
        self.fc_g2 = torch.nn.Linear(1500, output_dim)
        self.relu = nn.ReLU()
        self.dropout = nn.Dropout(dropout)
        # 단백질 서열 표현을 위한 1D Conv 층 구성
        self.embedding_xt = nn.Embedding(num_features_xt + 1, embed_dim)
        self.conv_xt_1 = nn.Conv1d(in_channels=1000, out_channels=n_filters, kernel_size=8)
        self.fc1_xt = nn.Linear(32*121, output_dim)
        # 층 결합
        self.fc1 = nn.Linear(256, 1024)
        self.fc2 = nn.Linear(1024, 512)
        self.out = nn.Linear(512, self.n_output)
    def forward(self, data):
        x, edge_index, batch = data.x, data.edge_index, data.batch
        target = data.target
        x = self.conv1(x, edge_index)
        x = self.relu(x)
        x = self.conv2(x, edge_index)
        x = self.relu(x)
        # gmp : global max pooling
        # gap : global mean pooling
        x = torch.cat([gmp(x, batch), gap(x, batch)], dim=1)
        x = self.relu(self.fc_g1(x))
        x = self.dropout(x)
        x = self.fc_g2(x)
        embedded_xt = self.embedding_xt(target)
        conv_xt = self.conv_xt_1(embedded_xt)
        # flatten
        xt = conv_xt.view(-1, 32 * 121)
        xt = self.fc1_xt(xt)
        # concat
        xc = torch.cat((x, xt), 1)
        # Dense 층 추가
        xc = self.fc1(xc)
        xc = self.relu(xc)
        xc = self.dropout(xc)
        xc = self.fc2(xc)
        xc = self.relu(xc)
        xc = self.dropout(xc)
        out = self.out(xc)
        return out
0.0s
Python
→

4-4. GIN

# GINConv model
class GINConvNet(torch.nn.Module):
    def __init__(self, n_output=1,num_features_xd=78, num_features_xt=25,
                 n_filters=32, embed_dim=128, output_dim=128, dropout=0.2):
        super(GINConvNet, self).__init__()
        dim = 32
        # 약물 분자 표현을 위한 GIN 층 구성
        self.dropout = nn.Dropout(dropout)
        self.relu = nn.ReLU()
        self.n_output = n_output
        nn1 = Sequential(Linear(num_features_xd, dim), ReLU(), Linear(dim, dim))
        self.conv1 = GINConv(nn1)
        self.bn1 = torch.nn.BatchNorm1d(dim)
        nn2 = Sequential(Linear(dim, dim), ReLU(), Linear(dim, dim))
        self.conv2 = GINConv(nn2)
        self.bn2 = torch.nn.BatchNorm1d(dim)
        nn3 = Sequential(Linear(dim, dim), ReLU(), Linear(dim, dim))
        self.conv3 = GINConv(nn3)
        self.bn3 = torch.nn.BatchNorm1d(dim)
        nn4 = Sequential(Linear(dim, dim), ReLU(), Linear(dim, dim))
        self.conv4 = GINConv(nn4)
        self.bn4 = torch.nn.BatchNorm1d(dim)
        nn5 = Sequential(Linear(dim, dim), ReLU(), Linear(dim, dim))
        self.conv5 = GINConv(nn5)
        self.bn5 = torch.nn.BatchNorm1d(dim)
        self.fc1_xd = Linear(dim, output_dim)
        # 단백질 서열 표현을 위한 1D Conv 층 구성
        self.embedding_xt = nn.Embedding(num_features_xt + 1, embed_dim)
        self.conv_xt_1 = nn.Conv1d(in_channels=1000, out_channels=n_filters, kernel_size=8)
        self.fc1_xt = nn.Linear(32*121, output_dim)
        # 층 결합
        self.fc1 = nn.Linear(256, 1024)
        self.fc2 = nn.Linear(1024, 256)
        self.out = nn.Linear(256, self.n_output)
    def forward(self, data):
        x, edge_index, batch = data.x, data.edge_index, data.batch
        target = data.target
        x = F.relu(self.conv1(x, edge_index))
        x = self.bn1(x)
        x = F.relu(self.conv2(x, edge_index))
        x = self.bn2(x)
        x = F.relu(self.conv3(x, edge_index))
        x = self.bn3(x)
        x = F.relu(self.conv4(x, edge_index))
        x = self.bn4(x)
        x = F.relu(self.conv5(x, edge_index))
        x = self.bn5(x)
        x = global_add_pool(x, batch)
        x = F.relu(self.fc1_xd(x))
        x = F.dropout(x, p=0.2, training=self.training)
        embedded_xt = self.embedding_xt(target)
        conv_xt = self.conv_xt_1(embedded_xt)
        # flatten
        xt = conv_xt.view(-1, 32 * 121)
        xt = self.fc1_xt(xt)
        # concat
        xc = torch.cat((x, xt), 1)
        # Dense 층 추가
        xc = self.fc1(xc)
        xc = self.relu(xc)
        xc = self.dropout(xc)
        xc = self.fc2(xc)
        xc = self.relu(xc)
        xc = self.dropout(xc)
        out = self.out(xc)
        return out
0.0s
Python
→

5. Training

5-1. Train

# training function at each epoch
def train(model, device, train_loader, optimizer, epoch):
    print('Training on {} samples...'.format(len(train_loader.dataset)))
    model.train()
    # train_loader에서 batch를 불러와 학습을 진행합니다.
    for batch_idx, data in enumerate(train_loader):
        # 학습에 사용할 데이터를 불러옵니다.
        data = data.to(device)
        # optimizer는 딥러닝 프로세스에서 실제 파라미터를 갱신시키는 부분입니다.
        # 오차 역전파 알고리즘과 같은 방식으로, 각 파라미터의 기울기를 이용하여 실제 가중치의 변화를 주는 부분입니다.
        # 역전파 단계 전에, 모델의 학습 가능한 가중치인 Optimizer 객체를 사용하여 갱신할 변수들에 대한 모든 변화도를 0으로 초기화합니다.
        # 이렇게 설정하는 이유는 기본적으로 역전파(loss.backward)를 호출할 때마다 변화도가 버퍼에 덮어쓰이지 않고 누적되기 때문입니다.
        # 더 자세한 내용은 torch.autograd.backward documentation을 참조하시면 됩니다.
        optimizer.zero_grad()
        # 순전파 단계 : 모델에 data를 전달하여 예상되는 output을 계산합니다.
        output = model(data)
        # 손실을 계산하고 출력합니다.
        loss = loss_fn(output, data.y.view(-1, 1).float().to(device))
        # 역전파 단계 : 모델의 매개변수에 대한 손실의 변화도를 계산합니다.
        loss.backward()
        # Optimizer 객체의 step 함수를 호출하면 매개변수가 갱신됩니다.
        optimizer.step()
        # 진행 상황을 출력합니다.
        if batch_idx % LOG_INTERVAL == 0:
            print('Train epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(epoch,
                                                                           batch_idx * len(data.x),
                                                                           len(train_loader.dataset),
                                                                           100. * batch_idx / len(train_loader),
                                                                           loss.item()))
0.0s
Python
→

Optimizer 참고자료https://bl.ocks.org/EmilienDupont/raw/aaf429be5705b219aaaf8d691e27ca87/

손실 최소화를 시작하려면 함수 히트 맵의 아무 곳이나 클릭하시면 됩니다. 하단 막대의 원을 클릭하면 다양한 알고리즘들의 시각화를 활성화/비활성화 할 수 있습니다. 전역 최소값은 왼쪽이 배치되어 있으며, 지역 최소값은 오른쪽에 배치되어 있습니다.

5-2. Predict

def predicting(model, device, loader):
    model.eval()
    # 학습을 통해 예측된 값을 저장합니다.
    total_preds = torch.Tensor()
    # 실제 값을 저장합니다.
    total_labels = torch.Tensor()
    print('Make prediction for {} samples...'.format(len(loader.dataset)))
    with torch.no_grad():
        for data in loader:
            data = data.to(device)
            output = model(data)
            total_preds = torch.cat((total_preds, output.cpu()), 0)
            total_labels = torch.cat((total_labels, data.y.view(-1, 1).cpu()), 0)
    
    # n차원 형태의 텐서를 1차원 형태로 평평하게 만들어 반환합니다.
    return total_labels.numpy().flatten(),total_preds.numpy().flatten()
0.0s
Python
→

5-3. Main

datasets = ['davis','kiba']
modeling = [GINConvNet, GATNet, GAT_GCN, GCNNet]
cuda_name = "cuda:0"
0.0s
Python
→

아래의 코드는 하이퍼 파라미터입니다.하이퍼 파라미터란 학습 프로세스 제어에 사용되는 값을 갖는 매개 변수입니다. 이와 반대로 다른 파라미터의 값은 훈련을 통해 도출됩니다.

TRAIN_BATCH_SIZE : 학습에 사용되는 데이터의 배치 크기입니다.

TEST_BATCH_SIZE : 테스트에 사용되는 데이터의 배치 크기입니다.

LR : Learning Rate로 학습율을 뜻합니다.

LOG_INTERVAL : log를 찍는 간격입니다.

NUM_EPOCHS : 훈련 반복 횟수입니다.

TRAIN_BATCH_SIZE = 512
TEST_BATCH_SIZE = 512
LR = 0.0005
LOG_INTERVAL = 20
NUM_EPOCHS = 1000
0.0s
Python
→
# Main program: iterate over different datasets
print('-----------------------------------------------------')
print('Select Dataset Number...')
print('|  0 : davis  |  1 : kiba  |')
d = int(input())
print('The ' + datasets[d] + ' dataset has been selected!')
print('-----------------------------------------------------')
print('Select Model Number...')
print('|  0 : GIN  |  1 : GAT  |  2 : GAT_GCN  |  3 : GCN  |')
m = int(input())
print('The ' + modeling[m].__name__ + ' model has been selected!')
print('-----------------------------------------------------')
print('\nrunning on ', modeling[m].__name__ + '_' + datasets[d])
processed_data_file_train = '/content/drive/My Drive/GraphDTA/data/processed/' + datasets[d] + '_train.pt'
processed_data_file_test = '/content/drive/My Drive/GraphDTA/data/processed/' + datasets[d] + '_test.pt'
if ((not os.path.isfile(processed_data_file_train)) or (not os.path.isfile(processed_data_file_test))):
    print('please run create_data.py to prepare data in pytorch format!')
else:
    train_data = TestbedDataset(root='/content/drive/My Drive/GraphDTA/data', dataset=datasets[d]+'_train')
    test_data = TestbedDataset(root='/content/drive/My Drive/GraphDTA/data', dataset=datasets[d]+'_test')
    
    # make data PyTorch mini-batch processing ready
    train_loader = DataLoader(train_data, batch_size=TRAIN_BATCH_SIZE, shuffle=True)
    test_loader = DataLoader(test_data, batch_size=TEST_BATCH_SIZE, shuffle=False)
    # training the model
    device = torch.device(cuda_name if torch.cuda.is_available() else "cpu")
    model = modeling[m]().to(device)
    # 모델의 손실함수로 MSE(Mean Square Error)를 사용합니다.
    loss_fn = nn.MSELoss()
    # 옵티마이저는 Adam을 사용합니다.
    optimizer = torch.optim.Adam(model.parameters(), lr=LR)
    best_mse = 1000
    best_ci = 0
    best_epoch = -1
    # 학습을 진행하면서 모델의 변화를 저장합니다.
    model_file_name = 'model_' + modeling[m].__name__ + '_' + datasets[d] +  '.model'
    result_file_name = 'result_' + modeling[m].__name__ + '_' + datasets[d] +  '.csv'
    for epoch in range(NUM_EPOCHS):
        train(model, device, train_loader, optimizer, epoch+1)
        # G : 실제 값
        # P : 모델 학습을 통해 예측한 값
        G,P = predicting(model, device, test_loader)
        # 다양한 성능평가지표를 저장합니다.
        ret = [rmse(G,P),mse(G,P),pearson(G,P),spearman(G,P),ci(G,P)]
        
        if ret[1]<best_mse:
            torch.save(model.state_dict(), model_file_name)
            with open(result_file_name,'w') as f:
                f.write(','.join(map(str,ret)))
            best_epoch = epoch+1
            best_mse = ret[1]
            best_ci = ret[-1]
            print('rmse improved at epoch ', best_epoch, '; best_mse,best_ci:', best_mse,best_ci,modeling[m],datasets[d])
        else:
            print(ret[1],'No improvement since epoch ', best_epoch, '; best_mse,best_ci:', best_mse,best_ci,modeling[m],datasets[d])
Shift+Enter to run
Python
→
스트리밍 출력 내용이 길어서 마지막 5000줄이 삭제되었습니다. 0.32431558 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 168 [0/25046 (0%)] Loss: 0.217329 Train epoch: 168 [325280/25046 (41%)] Loss: 0.218972 Train epoch: 168 [659040/25046 (82%)] Loss: 0.248648 Make prediction for 5010 samples... 0.32635558 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 169 [0/25046 (0%)] Loss: 0.170223 Train epoch: 169 [327900/25046 (41%)] Loss: 0.219116 Train epoch: 169 [657240/25046 (82%)] Loss: 0.243762 Make prediction for 5010 samples... 0.3192075 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 170 [0/25046 (0%)] Loss: 0.159238 Train epoch: 170 [329240/25046 (41%)] Loss: 0.183609 Train epoch: 170 [648840/25046 (82%)] Loss: 0.287235 Make prediction for 5010 samples... 0.3268813 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 171 [0/25046 (0%)] Loss: 0.206227 Train epoch: 171 [331360/25046 (41%)] Loss: 0.222992 Train epoch: 171 [645400/25046 (82%)] Loss: 0.176406 Make prediction for 5010 samples... 0.31842497 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 172 [0/25046 (0%)] Loss: 0.191043 Train epoch: 172 [326520/25046 (41%)] Loss: 0.210799 Train epoch: 172 [658160/25046 (82%)] Loss: 0.203208 Make prediction for 5010 samples... 0.3747821 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 173 [0/25046 (0%)] Loss: 0.245716 Train epoch: 173 [323920/25046 (41%)] Loss: 0.187424 Train epoch: 173 [658320/25046 (82%)] Loss: 0.214630 Make prediction for 5010 samples... 0.32202744 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 174 [0/25046 (0%)] Loss: 0.213998 Train epoch: 174 [324140/25046 (41%)] Loss: 0.201029 Train epoch: 174 [665320/25046 (82%)] Loss: 0.245210 Make prediction for 5010 samples... 0.31329033 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 175 [0/25046 (0%)] Loss: 0.190844 Train epoch: 175 [328920/25046 (41%)] Loss: 0.257894 Train epoch: 175 [641960/25046 (82%)] Loss: 0.232822 Make prediction for 5010 samples... 0.31958085 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 176 [0/25046 (0%)] Loss: 0.178021 Train epoch: 176 [329980/25046 (41%)] Loss: 0.240212 Train epoch: 176 [654880/25046 (82%)] Loss: 0.201986 Make prediction for 5010 samples... 0.31805372 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 177 [0/25046 (0%)] Loss: 0.163982 Train epoch: 177 [325960/25046 (41%)] Loss: 0.201837 Train epoch: 177 [659200/25046 (82%)] Loss: 0.206936 Make prediction for 5010 samples... 0.32528254 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 178 [0/25046 (0%)] Loss: 0.228296 Train epoch: 178 [326480/25046 (41%)] Loss: 0.210811 Train epoch: 178 [654400/25046 (82%)] Loss: 0.202534 Make prediction for 5010 samples... 0.3401031 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 179 [0/25046 (0%)] Loss: 0.194066 Train epoch: 179 [323580/25046 (41%)] Loss: 0.235797 Train epoch: 179 [656840/25046 (82%)] Loss: 0.207586 Make prediction for 5010 samples... 0.3293649 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 180 [0/25046 (0%)] Loss: 0.212914 Train epoch: 180 [328560/25046 (41%)] Loss: 0.252533 Train epoch: 180 [646840/25046 (82%)] Loss: 0.208546 Make prediction for 5010 samples... 0.33443263 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 181 [0/25046 (0%)] Loss: 0.272412 Train epoch: 181 [330160/25046 (41%)] Loss: 0.212382 Train epoch: 181 [660920/25046 (82%)] Loss: 0.173525 Make prediction for 5010 samples... 0.32465878 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 182 [0/25046 (0%)] Loss: 0.218376 Train epoch: 182 [334360/25046 (41%)] Loss: 0.218888 Train epoch: 182 [663600/25046 (82%)] Loss: 0.223529 Make prediction for 5010 samples... 0.32748044 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 183 [0/25046 (0%)] Loss: 0.172697 Train epoch: 183 [324200/25046 (41%)] Loss: 0.217196 Train epoch: 183 [651680/25046 (82%)] Loss: 0.302814 Make prediction for 5010 samples... 0.32342857 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 184 [0/25046 (0%)] Loss: 0.230467 Train epoch: 184 [327960/25046 (41%)] Loss: 0.180403 Train epoch: 184 [661520/25046 (82%)] Loss: 0.220475 Make prediction for 5010 samples... 0.33721223 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 185 [0/25046 (0%)] Loss: 0.233119 Train epoch: 185 [328040/25046 (41%)] Loss: 0.236323 Train epoch: 185 [659040/25046 (82%)] Loss: 0.176114 Make prediction for 5010 samples... 0.3343154 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 186 [0/25046 (0%)] Loss: 0.206025 Train epoch: 186 [325300/25046 (41%)] Loss: 0.257755 Train epoch: 186 [665080/25046 (82%)] Loss: 0.285471 Make prediction for 5010 samples... 0.39941546 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 187 [0/25046 (0%)] Loss: 0.217849 Train epoch: 187 [332180/25046 (41%)] Loss: 0.194466 Train epoch: 187 [657680/25046 (82%)] Loss: 0.189754 Make prediction for 5010 samples... 0.337389 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 188 [0/25046 (0%)] Loss: 0.162628 Train epoch: 188 [328360/25046 (41%)] Loss: 0.225633 Train epoch: 188 [656200/25046 (82%)] Loss: 0.184619 Make prediction for 5010 samples... 0.35673413 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 189 [0/25046 (0%)] Loss: 0.292743 Train epoch: 189 [329460/25046 (41%)] Loss: 0.204204 Train epoch: 189 [654560/25046 (82%)] Loss: 0.193779 Make prediction for 5010 samples... 0.3357839 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 190 [0/25046 (0%)] Loss: 0.253419 Train epoch: 190 [330040/25046 (41%)] Loss: 0.200233 Train epoch: 190 [660840/25046 (82%)] Loss: 0.229302 Make prediction for 5010 samples... 0.3745217 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 191 [0/25046 (0%)] Loss: 0.190716 Train epoch: 191 [325820/25046 (41%)] Loss: 0.244172 Train epoch: 191 [664920/25046 (82%)] Loss: 0.230047 Make prediction for 5010 samples... 0.31854862 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 192 [0/25046 (0%)] Loss: 0.198490 Train epoch: 192 [326400/25046 (41%)] Loss: 0.205849 Train epoch: 192 [661840/25046 (82%)] Loss: 0.177192 Make prediction for 5010 samples... 0.32833737 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 193 [0/25046 (0%)] Loss: 0.167062 Train epoch: 193 [325320/25046 (41%)] Loss: 0.237889 Train epoch: 193 [651880/25046 (82%)] Loss: 0.203382 Make prediction for 5010 samples... 0.3490978 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 194 [0/25046 (0%)] Loss: 0.189889 Train epoch: 194 [327080/25046 (41%)] Loss: 0.198260 Train epoch: 194 [657240/25046 (82%)] Loss: 0.201565 Make prediction for 5010 samples... 0.36198127 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 195 [0/25046 (0%)] Loss: 0.186964 Train epoch: 195 [323380/25046 (41%)] Loss: 0.232525 Train epoch: 195 [656680/25046 (82%)] Loss: 0.181212 Make prediction for 5010 samples... 0.32486415 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 196 [0/25046 (0%)] Loss: 0.217010 Train epoch: 196 [324840/25046 (41%)] Loss: 0.238361 Train epoch: 196 [662000/25046 (82%)] Loss: 0.214913 Make prediction for 5010 samples... 0.32445666 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 197 [0/25046 (0%)] Loss: 0.186185 Train epoch: 197 [327160/25046 (41%)] Loss: 0.199873 Train epoch: 197 [652720/25046 (82%)] Loss: 0.207340 Make prediction for 5010 samples... 0.34656906 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 198 [0/25046 (0%)] Loss: 0.203706 Train epoch: 198 [326180/25046 (41%)] Loss: 0.262493 Train epoch: 198 [668080/25046 (82%)] Loss: 0.230588 Make prediction for 5010 samples... 0.34589255 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 199 [0/25046 (0%)] Loss: 0.222006 Train epoch: 199 [331840/25046 (41%)] Loss: 0.200042 Train epoch: 199 [652680/25046 (82%)] Loss: 0.213210 Make prediction for 5010 samples... 0.36608046 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 200 [0/25046 (0%)] Loss: 0.250228 Train epoch: 200 [322440/25046 (41%)] Loss: 0.145883 Train epoch: 200 [653080/25046 (82%)] Loss: 0.236781 Make prediction for 5010 samples... 0.32593438 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 201 [0/25046 (0%)] Loss: 0.227003 Train epoch: 201 [330660/25046 (41%)] Loss: 0.176324 Train epoch: 201 [659760/25046 (82%)] Loss: 0.199973 Make prediction for 5010 samples... 0.31987342 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 202 [0/25046 (0%)] Loss: 0.221687 Train epoch: 202 [328140/25046 (41%)] Loss: 0.221406 Train epoch: 202 [657640/25046 (82%)] Loss: 0.240925 Make prediction for 5010 samples... 0.35607556 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 203 [0/25046 (0%)] Loss: 0.213615 Train epoch: 203 [333360/25046 (41%)] Loss: 0.228745 Train epoch: 203 [639480/25046 (82%)] Loss: 0.172875 Make prediction for 5010 samples... 0.34908175 No improvement since epoch 142 ; best_mse,best_ci: 0.31239757 0.853386350698445 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 204 [0/25046 (0%)] Loss: 0.189913 Train epoch: 204 [328780/25046 (41%)] Loss: 0.190170 Train epoch: 204 [659960/25046 (82%)] Loss: 0.227911 Make prediction for 5010 samples... rmse improved at epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 205 [0/25046 (0%)] Loss: 0.212563 Train epoch: 205 [331780/25046 (41%)] Loss: 0.237334 Train epoch: 205 [657640/25046 (82%)] Loss: 0.237509 Make prediction for 5010 samples... 0.32531053 No improvement since epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 206 [0/25046 (0%)] Loss: 0.187604 Train epoch: 206 [324780/25046 (41%)] Loss: 0.191747 Train epoch: 206 [651320/25046 (82%)] Loss: 0.235952 Make prediction for 5010 samples... 0.3410831 No improvement since epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 207 [0/25046 (0%)] Loss: 0.193026 Train epoch: 207 [327740/25046 (41%)] Loss: 0.231622 Train epoch: 207 [652920/25046 (82%)] Loss: 0.183641 Make prediction for 5010 samples... 0.37284258 No improvement since epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 208 [0/25046 (0%)] Loss: 0.190674 Train epoch: 208 [324200/25046 (41%)] Loss: 0.188397 Train epoch: 208 [655480/25046 (82%)] Loss: 0.186833 Make prediction for 5010 samples... 0.32202742 No improvement since epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 209 [0/25046 (0%)] Loss: 0.228437 Train epoch: 209 [326160/25046 (41%)] Loss: 0.263491 Train epoch: 209 [657040/25046 (82%)] Loss: 0.177416 Make prediction for 5010 samples... 0.34670502 No improvement since epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 210 [0/25046 (0%)] Loss: 0.181124 Train epoch: 210 [325920/25046 (41%)] Loss: 0.168254 Train epoch: 210 [654520/25046 (82%)] Loss: 0.182635 Make prediction for 5010 samples... 0.32977676 No improvement since epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 211 [0/25046 (0%)] Loss: 0.152289 Train epoch: 211 [327880/25046 (41%)] Loss: 0.190588 Train epoch: 211 [663080/25046 (82%)] Loss: 0.207443 Make prediction for 5010 samples... 0.35529178 No improvement since epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 212 [0/25046 (0%)] Loss: 0.233184 Train epoch: 212 [329480/25046 (41%)] Loss: 0.224433 Train epoch: 212 [654520/25046 (82%)] Loss: 0.264130 Make prediction for 5010 samples... 0.35755813 No improvement since epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 213 [0/25046 (0%)] Loss: 0.198610 Train epoch: 213 [325540/25046 (41%)] Loss: 0.180388 Train epoch: 213 [654240/25046 (82%)] Loss: 0.157889 Make prediction for 5010 samples... 0.31881934 No improvement since epoch 204 ; best_mse,best_ci: 0.3087972 0.8569499832527572 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 214 [0/25046 (0%)] Loss: 0.216600 Train epoch: 214 [327460/25046 (41%)] Loss: 0.199096 Train epoch: 214 [659160/25046 (82%)] Loss: 0.173220 Make prediction for 5010 samples... rmse improved at epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 215 [0/25046 (0%)] Loss: 0.188735 Train epoch: 215 [326440/25046 (41%)] Loss: 0.192885 Train epoch: 215 [662600/25046 (82%)] Loss: 0.171881 Make prediction for 5010 samples... 0.33095965 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 216 [0/25046 (0%)] Loss: 0.239321 Train epoch: 216 [324980/25046 (41%)] Loss: 0.142530 Train epoch: 216 [651480/25046 (82%)] Loss: 0.245361 Make prediction for 5010 samples... 0.34394187 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 217 [0/25046 (0%)] Loss: 0.206081 Train epoch: 217 [327060/25046 (41%)] Loss: 0.186840 Train epoch: 217 [657600/25046 (82%)] Loss: 0.214766 Make prediction for 5010 samples... 0.33072346 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 218 [0/25046 (0%)] Loss: 0.162627 Train epoch: 218 [329120/25046 (41%)] Loss: 0.207072 Train epoch: 218 [651000/25046 (82%)] Loss: 0.206841 Make prediction for 5010 samples... 0.31827876 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 219 [0/25046 (0%)] Loss: 0.192460 Train epoch: 219 [329780/25046 (41%)] Loss: 0.175977 Train epoch: 219 [654600/25046 (82%)] Loss: 0.170765 Make prediction for 5010 samples... 0.33954835 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 220 [0/25046 (0%)] Loss: 0.210005 Train epoch: 220 [321880/25046 (41%)] Loss: 0.188634 Train epoch: 220 [651720/25046 (82%)] Loss: 0.206837 Make prediction for 5010 samples... 0.3364092 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 221 [0/25046 (0%)] Loss: 0.194555 Train epoch: 221 [331140/25046 (41%)] Loss: 0.186361 Train epoch: 221 [657160/25046 (82%)] Loss: 0.208353 Make prediction for 5010 samples... 0.33288848 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 222 [0/25046 (0%)] Loss: 0.166279 Train epoch: 222 [327880/25046 (41%)] Loss: 0.219247 Train epoch: 222 [656240/25046 (82%)] Loss: 0.197767 Make prediction for 5010 samples... 0.32766166 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 223 [0/25046 (0%)] Loss: 0.195465 Train epoch: 223 [331240/25046 (41%)] Loss: 0.167877 Train epoch: 223 [649960/25046 (82%)] Loss: 0.172592 Make prediction for 5010 samples... 0.32358208 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 224 [0/25046 (0%)] Loss: 0.170348 Train epoch: 224 [329120/25046 (41%)] Loss: 0.216703 Train epoch: 224 [662800/25046 (82%)] Loss: 0.188107 Make prediction for 5010 samples... 0.33764777 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 225 [0/25046 (0%)] Loss: 0.181292 Train epoch: 225 [324420/25046 (41%)] Loss: 0.196331 Train epoch: 225 [656560/25046 (82%)] Loss: 0.192243 Make prediction for 5010 samples... 0.32747045 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 226 [0/25046 (0%)] Loss: 0.163479 Train epoch: 226 [328900/25046 (41%)] Loss: 0.182361 Train epoch: 226 [648520/25046 (82%)] Loss: 0.159691 Make prediction for 5010 samples... 0.3285148 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 227 [0/25046 (0%)] Loss: 0.237804 Train epoch: 227 [327240/25046 (41%)] Loss: 0.175086 Train epoch: 227 [651440/25046 (82%)] Loss: 0.177854 Make prediction for 5010 samples... 0.33434036 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 228 [0/25046 (0%)] Loss: 0.196509 Train epoch: 228 [328000/25046 (41%)] Loss: 0.182437 Train epoch: 228 [655160/25046 (82%)] Loss: 0.215088 Make prediction for 5010 samples... 0.33868843 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 229 [0/25046 (0%)] Loss: 0.181888 Train epoch: 229 [331560/25046 (41%)] Loss: 0.172614 Train epoch: 229 [647360/25046 (82%)] Loss: 0.329993 Make prediction for 5010 samples... 0.387388 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 230 [0/25046 (0%)] Loss: 0.225965 Train epoch: 230 [331440/25046 (41%)] Loss: 0.222857 Train epoch: 230 [664720/25046 (82%)] Loss: 0.217101 Make prediction for 5010 samples... 0.31740192 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 231 [0/25046 (0%)] Loss: 0.157672 Train epoch: 231 [330600/25046 (41%)] Loss: 0.177973 Train epoch: 231 [657080/25046 (82%)] Loss: 0.172250 Make prediction for 5010 samples... 0.3130499 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 232 [0/25046 (0%)] Loss: 0.191616 Train epoch: 232 [332240/25046 (41%)] Loss: 0.207994 Train epoch: 232 [655320/25046 (82%)] Loss: 0.175885 Make prediction for 5010 samples... 0.30984256 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 233 [0/25046 (0%)] Loss: 0.171333 Train epoch: 233 [331900/25046 (41%)] Loss: 0.176440 Train epoch: 233 [658720/25046 (82%)] Loss: 0.227612 Make prediction for 5010 samples... 0.3517062 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 234 [0/25046 (0%)] Loss: 0.196817 Train epoch: 234 [329560/25046 (41%)] Loss: 0.153031 Train epoch: 234 [648280/25046 (82%)] Loss: 0.139262 Make prediction for 5010 samples... 0.3290919 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 235 [0/25046 (0%)] Loss: 0.193408 Train epoch: 235 [325260/25046 (41%)] Loss: 0.163098 Train epoch: 235 [654320/25046 (82%)] Loss: 0.186195 Make prediction for 5010 samples... 0.32148093 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 236 [0/25046 (0%)] Loss: 0.151398 Train epoch: 236 [323740/25046 (41%)] Loss: 0.178855 Train epoch: 236 [660840/25046 (82%)] Loss: 0.203970 Make prediction for 5010 samples... 0.3232515 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 237 [0/25046 (0%)] Loss: 0.214631 Train epoch: 237 [328520/25046 (41%)] Loss: 0.194380 Train epoch: 237 [646000/25046 (82%)] Loss: 0.217287 Make prediction for 5010 samples... 0.33432004 No improvement since epoch 214 ; best_mse,best_ci: 0.30795443 0.8559309054859741 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 238 [0/25046 (0%)] Loss: 0.139529 Train epoch: 238 [329820/25046 (41%)] Loss: 0.232859 Train epoch: 238 [652600/25046 (82%)] Loss: 0.174378 Make prediction for 5010 samples... rmse improved at epoch 238 ; best_mse,best_ci: 0.30687955 0.8579848357430117 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 239 [0/25046 (0%)] Loss: 0.177889 Train epoch: 239 [329180/25046 (41%)] Loss: 0.152916 Train epoch: 239 [660400/25046 (82%)] Loss: 0.142207 Make prediction for 5010 samples... 0.314676 No improvement since epoch 238 ; best_mse,best_ci: 0.30687955 0.8579848357430117 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 240 [0/25046 (0%)] Loss: 0.151702 Train epoch: 240 [329280/25046 (41%)] Loss: 0.192803 Train epoch: 240 [657720/25046 (82%)] Loss: 0.268140 Make prediction for 5010 samples... 0.31686458 No improvement since epoch 238 ; best_mse,best_ci: 0.30687955 0.8579848357430117 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 241 [0/25046 (0%)] Loss: 0.193548 Train epoch: 241 [329540/25046 (41%)] Loss: 0.164824 Train epoch: 241 [656360/25046 (82%)] Loss: 0.195685 Make prediction for 5010 samples... 0.32141334 No improvement since epoch 238 ; best_mse,best_ci: 0.30687955 0.8579848357430117 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 242 [0/25046 (0%)] Loss: 0.166217 Train epoch: 242 [331440/25046 (41%)] Loss: 0.172731 Train epoch: 242 [649360/25046 (82%)] Loss: 0.177149 Make prediction for 5010 samples... 0.35077092 No improvement since epoch 238 ; best_mse,best_ci: 0.30687955 0.8579848357430117 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 243 [0/25046 (0%)] Loss: 0.174817 Train epoch: 243 [329320/25046 (41%)] Loss: 0.164079 Train epoch: 243 [652280/25046 (82%)] Loss: 0.148316 Make prediction for 5010 samples... 0.33995152 No improvement since epoch 238 ; best_mse,best_ci: 0.30687955 0.8579848357430117 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 244 [0/25046 (0%)] Loss: 0.164439 Train epoch: 244 [332320/25046 (41%)] Loss: 0.200588 Train epoch: 244 [653720/25046 (82%)] Loss: 0.237278 Make prediction for 5010 samples... 0.31738767 No improvement since epoch 238 ; best_mse,best_ci: 0.30687955 0.8579848357430117 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 245 [0/25046 (0%)] Loss: 0.173168 Train epoch: 245 [324460/25046 (41%)] Loss: 0.149027 Train epoch: 245 [650800/25046 (82%)] Loss: 0.167447 Make prediction for 5010 samples... 0.32014477 No improvement since epoch 238 ; best_mse,best_ci: 0.30687955 0.8579848357430117 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 246 [0/25046 (0%)] Loss: 0.204177 Train epoch: 246 [334940/25046 (41%)] Loss: 0.199856 Train epoch: 246 [658560/25046 (82%)] Loss: 0.195015 Make prediction for 5010 samples... rmse improved at epoch 246 ; best_mse,best_ci: 0.30600083 0.8616406661956713 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 247 [0/25046 (0%)] Loss: 0.179737 Train epoch: 247 [324940/25046 (41%)] Loss: 0.179875 Train epoch: 247 [652800/25046 (82%)] Loss: 0.202086 Make prediction for 5010 samples... 0.3123676 No improvement since epoch 246 ; best_mse,best_ci: 0.30600083 0.8616406661956713 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 248 [0/25046 (0%)] Loss: 0.167685 Train epoch: 248 [325560/25046 (41%)] Loss: 0.179464 Train epoch: 248 [659720/25046 (82%)] Loss: 0.147450 Make prediction for 5010 samples... 0.31494236 No improvement since epoch 246 ; best_mse,best_ci: 0.30600083 0.8616406661956713 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 249 [0/25046 (0%)] Loss: 0.188253 Train epoch: 249 [327480/25046 (41%)] Loss: 0.163509 Train epoch: 249 [653720/25046 (82%)] Loss: 0.226091 Make prediction for 5010 samples... 0.36146468 No improvement since epoch 246 ; best_mse,best_ci: 0.30600083 0.8616406661956713 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 250 [0/25046 (0%)] Loss: 0.163226 Train epoch: 250 [326480/25046 (41%)] Loss: 0.189892 Train epoch: 250 [653920/25046 (82%)] Loss: 0.169621 Make prediction for 5010 samples... 0.34305036 No improvement since epoch 246 ; best_mse,best_ci: 0.30600083 0.8616406661956713 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 251 [0/25046 (0%)] Loss: 0.209922 Train epoch: 251 [329480/25046 (41%)] Loss: 0.258084 Train epoch: 251 [657400/25046 (82%)] Loss: 0.235205 Make prediction for 5010 samples... 0.31300277 No improvement since epoch 246 ; best_mse,best_ci: 0.30600083 0.8616406661956713 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 252 [0/25046 (0%)] Loss: 0.165107 Train epoch: 252 [330860/25046 (41%)] Loss: 0.133582 Train epoch: 252 [651880/25046 (82%)] Loss: 0.168787 Make prediction for 5010 samples... 0.3196965 No improvement since epoch 246 ; best_mse,best_ci: 0.30600083 0.8616406661956713 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 253 [0/25046 (0%)] Loss: 0.187376 Train epoch: 253 [329400/25046 (41%)] Loss: 0.205724 Train epoch: 253 [660920/25046 (82%)] Loss: 0.169149 Make prediction for 5010 samples... 0.30634472 No improvement since epoch 246 ; best_mse,best_ci: 0.30600083 0.8616406661956713 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 254 [0/25046 (0%)] Loss: 0.179483 Train epoch: 254 [327160/25046 (41%)] Loss: 0.184159 Train epoch: 254 [647680/25046 (82%)] Loss: 0.191498 Make prediction for 5010 samples... rmse improved at epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 255 [0/25046 (0%)] Loss: 0.165475 Train epoch: 255 [329260/25046 (41%)] Loss: 0.162073 Train epoch: 255 [655160/25046 (82%)] Loss: 0.180550 Make prediction for 5010 samples... 0.3400495 No improvement since epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 256 [0/25046 (0%)] Loss: 0.164321 Train epoch: 256 [328780/25046 (41%)] Loss: 0.211639 Train epoch: 256 [649920/25046 (82%)] Loss: 0.162887 Make prediction for 5010 samples... 0.3373687 No improvement since epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 257 [0/25046 (0%)] Loss: 0.172738 Train epoch: 257 [334840/25046 (41%)] Loss: 0.341231 Train epoch: 257 [654640/25046 (82%)] Loss: 0.209377 Make prediction for 5010 samples... 0.3179038 No improvement since epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 258 [0/25046 (0%)] Loss: 0.134100 Train epoch: 258 [328120/25046 (41%)] Loss: 0.184866 Train epoch: 258 [648160/25046 (82%)] Loss: 0.216858 Make prediction for 5010 samples... 0.32165518 No improvement since epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 259 [0/25046 (0%)] Loss: 0.185758 Train epoch: 259 [330020/25046 (41%)] Loss: 0.217370 Train epoch: 259 [646640/25046 (82%)] Loss: 0.176243 Make prediction for 5010 samples... 0.3992056 No improvement since epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 260 [0/25046 (0%)] Loss: 0.222698 Train epoch: 260 [330660/25046 (41%)] Loss: 0.168354 Train epoch: 260 [651760/25046 (82%)] Loss: 0.158944 Make prediction for 5010 samples... 0.31565884 No improvement since epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 261 [0/25046 (0%)] Loss: 0.182452 Train epoch: 261 [327140/25046 (41%)] Loss: 0.184063 Train epoch: 261 [666760/25046 (82%)] Loss: 0.221078 Make prediction for 5010 samples... 0.3149293 No improvement since epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 262 [0/25046 (0%)] Loss: 0.220208 Train epoch: 262 [328500/25046 (41%)] Loss: 0.190969 Train epoch: 262 [658080/25046 (82%)] Loss: 0.179716 Make prediction for 5010 samples... 0.33039552 No improvement since epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 263 [0/25046 (0%)] Loss: 0.149575 Train epoch: 263 [329240/25046 (41%)] Loss: 0.169811 Train epoch: 263 [652760/25046 (82%)] Loss: 0.136627 Make prediction for 5010 samples... 0.31206444 No improvement since epoch 254 ; best_mse,best_ci: 0.3043787 0.8543381944496559 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 264 [0/25046 (0%)] Loss: 0.167812 Train epoch: 264 [323260/25046 (41%)] Loss: 0.167777 Train epoch: 264 [660200/25046 (82%)] Loss: 0.213334 Make prediction for 5010 samples... rmse improved at epoch 264 ; best_mse,best_ci: 0.30202633 0.8470877946063616 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 265 [0/25046 (0%)] Loss: 0.166811 Train epoch: 265 [325020/25046 (41%)] Loss: 0.149655 Train epoch: 265 [658800/25046 (82%)] Loss: 0.171794 Make prediction for 5010 samples... 0.336775 No improvement since epoch 264 ; best_mse,best_ci: 0.30202633 0.8470877946063616 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 266 [0/25046 (0%)] Loss: 0.188562 Train epoch: 266 [327540/25046 (41%)] Loss: 0.139836 Train epoch: 266 [645880/25046 (82%)] Loss: 0.201722 Make prediction for 5010 samples... 0.32005793 No improvement since epoch 264 ; best_mse,best_ci: 0.30202633 0.8470877946063616 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 267 [0/25046 (0%)] Loss: 0.167041 Train epoch: 267 [335120/25046 (41%)] Loss: 0.223150 Train epoch: 267 [651440/25046 (82%)] Loss: 0.146557 Make prediction for 5010 samples... 0.31829983 No improvement since epoch 264 ; best_mse,best_ci: 0.30202633 0.8470877946063616 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 268 [0/25046 (0%)] Loss: 0.147828 Train epoch: 268 [328400/25046 (41%)] Loss: 0.172456 Train epoch: 268 [654560/25046 (82%)] Loss: 0.211419 Make prediction for 5010 samples... 0.32236022 No improvement since epoch 264 ; best_mse,best_ci: 0.30202633 0.8470877946063616 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 269 [0/25046 (0%)] Loss: 0.152525 Train epoch: 269 [329060/25046 (41%)] Loss: 0.171482 Train epoch: 269 [658360/25046 (82%)] Loss: 0.140071 Make prediction for 5010 samples... 0.30261606 No improvement since epoch 264 ; best_mse,best_ci: 0.30202633 0.8470877946063616 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 270 [0/25046 (0%)] Loss: 0.165563 Train epoch: 270 [331180/25046 (41%)] Loss: 0.179034 Train epoch: 270 [664120/25046 (82%)] Loss: 0.217510 Make prediction for 5010 samples... 0.32517782 No improvement since epoch 264 ; best_mse,best_ci: 0.30202633 0.8470877946063616 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 271 [0/25046 (0%)] Loss: 0.142777 Train epoch: 271 [329440/25046 (41%)] Loss: 0.190462 Train epoch: 271 [646720/25046 (82%)] Loss: 0.175725 Make prediction for 5010 samples... rmse improved at epoch 271 ; best_mse,best_ci: 0.29961443 0.8632705167508419 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 272 [0/25046 (0%)] Loss: 0.151097 Train epoch: 272 [324560/25046 (41%)] Loss: 0.155266 Train epoch: 272 [666600/25046 (82%)] Loss: 0.154077 Make prediction for 5010 samples... 0.33153188 No improvement since epoch 271 ; best_mse,best_ci: 0.29961443 0.8632705167508419 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 273 [0/25046 (0%)] Loss: 0.211776 Train epoch: 273 [327880/25046 (41%)] Loss: 0.174423 Train epoch: 273 [655000/25046 (82%)] Loss: 0.183350 Make prediction for 5010 samples... 0.31100774 No improvement since epoch 271 ; best_mse,best_ci: 0.29961443 0.8632705167508419 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 274 [0/25046 (0%)] Loss: 0.146748 Train epoch: 274 [322540/25046 (41%)] Loss: 0.185645 Train epoch: 274 [660360/25046 (82%)] Loss: 0.147886 Make prediction for 5010 samples... 0.31315476 No improvement since epoch 271 ; best_mse,best_ci: 0.29961443 0.8632705167508419 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 275 [0/25046 (0%)] Loss: 0.167431 Train epoch: 275 [324880/25046 (41%)] Loss: 0.192790 Train epoch: 275 [661840/25046 (82%)] Loss: 0.143073 Make prediction for 5010 samples... 0.3042666 No improvement since epoch 271 ; best_mse,best_ci: 0.29961443 0.8632705167508419 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 276 [0/25046 (0%)] Loss: 0.150747 Train epoch: 276 [323740/25046 (41%)] Loss: 0.145413 Train epoch: 276 [654680/25046 (82%)] Loss: 0.130223 Make prediction for 5010 samples... rmse improved at epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 277 [0/25046 (0%)] Loss: 0.151180 Train epoch: 277 [327680/25046 (41%)] Loss: 0.170128 Train epoch: 277 [664160/25046 (82%)] Loss: 0.212776 Make prediction for 5010 samples... 0.30061772 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 278 [0/25046 (0%)] Loss: 0.137677 Train epoch: 278 [331820/25046 (41%)] Loss: 0.157251 Train epoch: 278 [662120/25046 (82%)] Loss: 0.158873 Make prediction for 5010 samples... 0.3260152 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 279 [0/25046 (0%)] Loss: 0.186663 Train epoch: 279 [326100/25046 (41%)] Loss: 0.140363 Train epoch: 279 [663360/25046 (82%)] Loss: 0.167889 Make prediction for 5010 samples... 0.32349685 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 280 [0/25046 (0%)] Loss: 0.167749 Train epoch: 280 [332100/25046 (41%)] Loss: 0.162944 Train epoch: 280 [655160/25046 (82%)] Loss: 0.171157 Make prediction for 5010 samples... 0.3224195 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 281 [0/25046 (0%)] Loss: 0.150017 Train epoch: 281 [332140/25046 (41%)] Loss: 0.222254 Train epoch: 281 [659000/25046 (82%)] Loss: 0.161869 Make prediction for 5010 samples... 0.322801 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 282 [0/25046 (0%)] Loss: 0.148762 Train epoch: 282 [326720/25046 (41%)] Loss: 0.135432 Train epoch: 282 [662840/25046 (82%)] Loss: 0.168047 Make prediction for 5010 samples... 0.29825607 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 283 [0/25046 (0%)] Loss: 0.130132 Train epoch: 283 [332420/25046 (41%)] Loss: 0.200547 Train epoch: 283 [659560/25046 (82%)] Loss: 0.170104 Make prediction for 5010 samples... 0.3834501 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 284 [0/25046 (0%)] Loss: 0.199140 Train epoch: 284 [326720/25046 (41%)] Loss: 0.177227 Train epoch: 284 [660280/25046 (82%)] Loss: 0.168207 Make prediction for 5010 samples... 0.30192608 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 285 [0/25046 (0%)] Loss: 0.148157 Train epoch: 285 [328200/25046 (41%)] Loss: 0.190320 Train epoch: 285 [658320/25046 (82%)] Loss: 0.162980 Make prediction for 5010 samples... 0.32501855 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 286 [0/25046 (0%)] Loss: 0.136943 Train epoch: 286 [330580/25046 (41%)] Loss: 0.121905 Train epoch: 286 [649680/25046 (82%)] Loss: 0.183029 Make prediction for 5010 samples... 0.34081927 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 287 [0/25046 (0%)] Loss: 0.195565 Train epoch: 287 [325000/25046 (41%)] Loss: 0.206559 Train epoch: 287 [658720/25046 (82%)] Loss: 0.176818 Make prediction for 5010 samples... 0.30339172 No improvement since epoch 276 ; best_mse,best_ci: 0.29810312 0.8590343143118246 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 288 [0/25046 (0%)] Loss: 0.151021 Train epoch: 288 [328780/25046 (41%)] Loss: 0.185260 Train epoch: 288 [664400/25046 (82%)] Loss: 0.167248 Make prediction for 5010 samples... rmse improved at epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 289 [0/25046 (0%)] Loss: 0.169435 Train epoch: 289 [328840/25046 (41%)] Loss: 0.172673 Train epoch: 289 [663200/25046 (82%)] Loss: 0.173236 Make prediction for 5010 samples... 0.29687136 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 290 [0/25046 (0%)] Loss: 0.181990 Train epoch: 290 [327860/25046 (41%)] Loss: 0.213545 Train epoch: 290 [653600/25046 (82%)] Loss: 0.142741 Make prediction for 5010 samples... 0.31128544 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 291 [0/25046 (0%)] Loss: 0.165152 Train epoch: 291 [329480/25046 (41%)] Loss: 0.127632 Train epoch: 291 [649240/25046 (82%)] Loss: 0.185926 Make prediction for 5010 samples... 0.3108026 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 292 [0/25046 (0%)] Loss: 0.150018 Train epoch: 292 [327620/25046 (41%)] Loss: 0.142524 Train epoch: 292 [653680/25046 (82%)] Loss: 0.142290 Make prediction for 5010 samples... 0.31374472 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 293 [0/25046 (0%)] Loss: 0.189930 Train epoch: 293 [325960/25046 (41%)] Loss: 0.163197 Train epoch: 293 [659160/25046 (82%)] Loss: 0.174006 Make prediction for 5010 samples... 0.29871017 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 294 [0/25046 (0%)] Loss: 0.152719 Train epoch: 294 [330320/25046 (41%)] Loss: 0.154097 Train epoch: 294 [658480/25046 (82%)] Loss: 0.189879 Make prediction for 5010 samples... 0.29790935 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 295 [0/25046 (0%)] Loss: 0.155752 Train epoch: 295 [328740/25046 (41%)] Loss: 0.170088 Train epoch: 295 [662720/25046 (82%)] Loss: 0.258512 Make prediction for 5010 samples... 0.36351287 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 296 [0/25046 (0%)] Loss: 0.208091 Train epoch: 296 [328800/25046 (41%)] Loss: 0.158615 Train epoch: 296 [659960/25046 (82%)] Loss: 0.177661 Make prediction for 5010 samples... 0.29898518 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 297 [0/25046 (0%)] Loss: 0.173242 Train epoch: 297 [327820/25046 (41%)] Loss: 0.158736 Train epoch: 297 [658640/25046 (82%)] Loss: 0.214708 Make prediction for 5010 samples... 0.29958278 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 298 [0/25046 (0%)] Loss: 0.175732 Train epoch: 298 [330400/25046 (41%)] Loss: 0.137541 Train epoch: 298 [656840/25046 (82%)] Loss: 0.162656 Make prediction for 5010 samples... 0.31934813 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 299 [0/25046 (0%)] Loss: 0.136196 Train epoch: 299 [324200/25046 (41%)] Loss: 0.166220 Train epoch: 299 [657200/25046 (82%)] Loss: 0.144597 Make prediction for 5010 samples... 0.301046 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 300 [0/25046 (0%)] Loss: 0.139140 Train epoch: 300 [330040/25046 (41%)] Loss: 0.150662 Train epoch: 300 [660880/25046 (82%)] Loss: 0.204515 Make prediction for 5010 samples... 0.33319986 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 301 [0/25046 (0%)] Loss: 0.178318 Train epoch: 301 [324700/25046 (41%)] Loss: 0.139898 Train epoch: 301 [647720/25046 (82%)] Loss: 0.170077 Make prediction for 5010 samples... 0.29775932 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 302 [0/25046 (0%)] Loss: 0.125584 Train epoch: 302 [332340/25046 (41%)] Loss: 0.184252 Train epoch: 302 [661000/25046 (82%)] Loss: 0.171103 Make prediction for 5010 samples... 0.32687232 No improvement since epoch 288 ; best_mse,best_ci: 0.29122147 0.8631341343115091 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 303 [0/25046 (0%)] Loss: 0.165166 Train epoch: 303 [331980/25046 (41%)] Loss: 0.170322 Train epoch: 303 [656400/25046 (82%)] Loss: 0.188746 Make prediction for 5010 samples... rmse improved at epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 304 [0/25046 (0%)] Loss: 0.155665 Train epoch: 304 [330200/25046 (41%)] Loss: 0.171736 Train epoch: 304 [658040/25046 (82%)] Loss: 0.175594 Make prediction for 5010 samples... 0.2936273 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 305 [0/25046 (0%)] Loss: 0.132575 Train epoch: 305 [333740/25046 (41%)] Loss: 0.150131 Train epoch: 305 [656040/25046 (82%)] Loss: 0.131954 Make prediction for 5010 samples... 0.29509965 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 306 [0/25046 (0%)] Loss: 0.136254 Train epoch: 306 [330640/25046 (41%)] Loss: 0.166067 Train epoch: 306 [661960/25046 (82%)] Loss: 0.146146 Make prediction for 5010 samples... 0.2978637 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 307 [0/25046 (0%)] Loss: 0.152554 Train epoch: 307 [327300/25046 (41%)] Loss: 0.132077 Train epoch: 307 [659120/25046 (82%)] Loss: 0.180321 Make prediction for 5010 samples... 0.306942 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 308 [0/25046 (0%)] Loss: 0.150859 Train epoch: 308 [328000/25046 (41%)] Loss: 0.125712 Train epoch: 308 [649280/25046 (82%)] Loss: 0.174356 Make prediction for 5010 samples... 0.29261535 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 309 [0/25046 (0%)] Loss: 0.113002 Train epoch: 309 [328540/25046 (41%)] Loss: 0.181012 Train epoch: 309 [651520/25046 (82%)] Loss: 0.176960 Make prediction for 5010 samples... 0.34242463 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 310 [0/25046 (0%)] Loss: 0.181456 Train epoch: 310 [334300/25046 (41%)] Loss: 0.151049 Train epoch: 310 [651000/25046 (82%)] Loss: 0.159787 Make prediction for 5010 samples... 0.312382 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 311 [0/25046 (0%)] Loss: 0.210275 Train epoch: 311 [325420/25046 (41%)] Loss: 0.149731 Train epoch: 311 [659840/25046 (82%)] Loss: 0.112316 Make prediction for 5010 samples... 0.3046815 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 312 [0/25046 (0%)] Loss: 0.135962 Train epoch: 312 [325740/25046 (41%)] Loss: 0.126522 Train epoch: 312 [658200/25046 (82%)] Loss: 0.163379 Make prediction for 5010 samples... 0.29837897 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 313 [0/25046 (0%)] Loss: 0.157145 Train epoch: 313 [325440/25046 (41%)] Loss: 0.207134 Train epoch: 313 [660160/25046 (82%)] Loss: 0.173482 Make prediction for 5010 samples... 0.4104772 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 314 [0/25046 (0%)] Loss: 0.222485 Train epoch: 314 [325780/25046 (41%)] Loss: 0.191096 Train epoch: 314 [661520/25046 (82%)] Loss: 0.183089 Make prediction for 5010 samples... 0.29089984 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 315 [0/25046 (0%)] Loss: 0.190057 Train epoch: 315 [326540/25046 (41%)] Loss: 0.152160 Train epoch: 315 [649240/25046 (82%)] Loss: 0.169380 Make prediction for 5010 samples... 0.29431722 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 316 [0/25046 (0%)] Loss: 0.161923 Train epoch: 316 [330640/25046 (41%)] Loss: 0.131373 Train epoch: 316 [645600/25046 (82%)] Loss: 0.134439 Make prediction for 5010 samples... 0.30785003 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 317 [0/25046 (0%)] Loss: 0.163860 Train epoch: 317 [326920/25046 (41%)] Loss: 0.185732 Train epoch: 317 [657200/25046 (82%)] Loss: 0.133579 Make prediction for 5010 samples... 0.3061291 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 318 [0/25046 (0%)] Loss: 0.161186 Train epoch: 318 [328660/25046 (41%)] Loss: 0.190840 Train epoch: 318 [653080/25046 (82%)] Loss: 0.152789 Make prediction for 5010 samples... 0.32388234 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 319 [0/25046 (0%)] Loss: 0.161058 Train epoch: 319 [329360/25046 (41%)] Loss: 0.164470 Train epoch: 319 [661080/25046 (82%)] Loss: 0.169237 Make prediction for 5010 samples... 0.3274839 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 320 [0/25046 (0%)] Loss: 0.160150 Train epoch: 320 [331760/25046 (41%)] Loss: 0.141148 Train epoch: 320 [659480/25046 (82%)] Loss: 0.179485 Make prediction for 5010 samples... 0.3127726 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 321 [0/25046 (0%)] Loss: 0.195597 Train epoch: 321 [329280/25046 (41%)] Loss: 0.161545 Train epoch: 321 [654800/25046 (82%)] Loss: 0.160381 Make prediction for 5010 samples... 0.3050009 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 322 [0/25046 (0%)] Loss: 0.165384 Train epoch: 322 [330460/25046 (41%)] Loss: 0.152867 Train epoch: 322 [660320/25046 (82%)] Loss: 0.139337 Make prediction for 5010 samples... 0.30073485 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 323 [0/25046 (0%)] Loss: 0.149000 Train epoch: 323 [326860/25046 (41%)] Loss: 0.155775 Train epoch: 323 [650400/25046 (82%)] Loss: 0.160882 Make prediction for 5010 samples... 0.29135427 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 324 [0/25046 (0%)] Loss: 0.158590 Train epoch: 324 [326860/25046 (41%)] Loss: 0.170949 Train epoch: 324 [653640/25046 (82%)] Loss: 0.218158 Make prediction for 5010 samples... 0.36779657 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 325 [0/25046 (0%)] Loss: 0.184669 Train epoch: 325 [328420/25046 (41%)] Loss: 0.160246 Train epoch: 325 [657480/25046 (82%)] Loss: 0.184003 Make prediction for 5010 samples... 0.29961282 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 326 [0/25046 (0%)] Loss: 0.146188 Train epoch: 326 [322380/25046 (41%)] Loss: 0.192146 Train epoch: 326 [652360/25046 (82%)] Loss: 0.163912 Make prediction for 5010 samples... 0.29684407 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 327 [0/25046 (0%)] Loss: 0.140595 Train epoch: 327 [327600/25046 (41%)] Loss: 0.184461 Train epoch: 327 [664480/25046 (82%)] Loss: 0.165062 Make prediction for 5010 samples... 0.30782062 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 328 [0/25046 (0%)] Loss: 0.158215 Train epoch: 328 [333220/25046 (41%)] Loss: 0.160639 Train epoch: 328 [659080/25046 (82%)] Loss: 0.153615 Make prediction for 5010 samples... 0.30306682 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 329 [0/25046 (0%)] Loss: 0.214057 Train epoch: 329 [329800/25046 (41%)] Loss: 0.185934 Train epoch: 329 [654600/25046 (82%)] Loss: 0.161088 Make prediction for 5010 samples... 0.29268378 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 330 [0/25046 (0%)] Loss: 0.161719 Train epoch: 330 [328040/25046 (41%)] Loss: 0.138936 Train epoch: 330 [660200/25046 (82%)] Loss: 0.183571 Make prediction for 5010 samples... 0.31060258 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 331 [0/25046 (0%)] Loss: 0.144245 Train epoch: 331 [328780/25046 (41%)] Loss: 0.140259 Train epoch: 331 [650200/25046 (82%)] Loss: 0.135536 Make prediction for 5010 samples... 0.31859824 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 332 [0/25046 (0%)] Loss: 0.130858 Train epoch: 332 [327720/25046 (41%)] Loss: 0.130893 Train epoch: 332 [651800/25046 (82%)] Loss: 0.142953 Make prediction for 5010 samples... 0.29786745 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 333 [0/25046 (0%)] Loss: 0.151972 Train epoch: 333 [331000/25046 (41%)] Loss: 0.151775 Train epoch: 333 [663240/25046 (82%)] Loss: 0.143601 Make prediction for 5010 samples... 0.30090556 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 334 [0/25046 (0%)] Loss: 0.163200 Train epoch: 334 [334460/25046 (41%)] Loss: 0.160760 Train epoch: 334 [648680/25046 (82%)] Loss: 0.156259 Make prediction for 5010 samples... 0.2893321 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 335 [0/25046 (0%)] Loss: 0.137454 Train epoch: 335 [328700/25046 (41%)] Loss: 0.138614 Train epoch: 335 [655800/25046 (82%)] Loss: 0.146445 Make prediction for 5010 samples... 0.30907872 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 336 [0/25046 (0%)] Loss: 0.127105 Train epoch: 336 [326400/25046 (41%)] Loss: 0.136154 Train epoch: 336 [662040/25046 (82%)] Loss: 0.142866 Make prediction for 5010 samples... 0.2893223 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 337 [0/25046 (0%)] Loss: 0.153484 Train epoch: 337 [329980/25046 (41%)] Loss: 0.205829 Train epoch: 337 [650960/25046 (82%)] Loss: 0.145399 Make prediction for 5010 samples... 0.3066598 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 338 [0/25046 (0%)] Loss: 0.164571 Train epoch: 338 [329060/25046 (41%)] Loss: 0.183343 Train epoch: 338 [654160/25046 (82%)] Loss: 0.192375 Make prediction for 5010 samples... 0.29712594 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 339 [0/25046 (0%)] Loss: 0.123817 Train epoch: 339 [332560/25046 (41%)] Loss: 0.133687 Train epoch: 339 [659320/25046 (82%)] Loss: 0.174707 Make prediction for 5010 samples... 0.29205397 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 340 [0/25046 (0%)] Loss: 0.127822 Train epoch: 340 [325700/25046 (41%)] Loss: 0.134107 Train epoch: 340 [648920/25046 (82%)] Loss: 0.141563 Make prediction for 5010 samples... 0.2962945 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 341 [0/25046 (0%)] Loss: 0.154807 Train epoch: 341 [327660/25046 (41%)] Loss: 0.124590 Train epoch: 341 [656920/25046 (82%)] Loss: 0.127485 Make prediction for 5010 samples... 0.30118307 No improvement since epoch 303 ; best_mse,best_ci: 0.2867403 0.863028995013809 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 342 [0/25046 (0%)] Loss: 0.143880 Train epoch: 342 [332220/25046 (41%)] Loss: 0.125741 Train epoch: 342 [645840/25046 (82%)] Loss: 0.170891 Make prediction for 5010 samples... rmse improved at epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 343 [0/25046 (0%)] Loss: 0.190199 Train epoch: 343 [332540/25046 (41%)] Loss: 0.149035 Train epoch: 343 [656480/25046 (82%)] Loss: 0.147461 Make prediction for 5010 samples... 0.30305615 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 344 [0/25046 (0%)] Loss: 0.135381 Train epoch: 344 [332300/25046 (41%)] Loss: 0.177271 Train epoch: 344 [652120/25046 (82%)] Loss: 0.135228 Make prediction for 5010 samples... 0.38973162 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 345 [0/25046 (0%)] Loss: 0.208817 Train epoch: 345 [322940/25046 (41%)] Loss: 0.163644 Train epoch: 345 [650080/25046 (82%)] Loss: 0.160277 Make prediction for 5010 samples... 0.2979173 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 346 [0/25046 (0%)] Loss: 0.150456 Train epoch: 346 [331320/25046 (41%)] Loss: 0.187567 Train epoch: 346 [658800/25046 (82%)] Loss: 0.122192 Make prediction for 5010 samples... 0.29516822 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 347 [0/25046 (0%)] Loss: 0.123936 Train epoch: 347 [328320/25046 (41%)] Loss: 0.139039 Train epoch: 347 [661880/25046 (82%)] Loss: 0.215092 Make prediction for 5010 samples... 0.3835492 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 348 [0/25046 (0%)] Loss: 0.211912 Train epoch: 348 [332460/25046 (41%)] Loss: 0.164328 Train epoch: 348 [662560/25046 (82%)] Loss: 0.142506 Make prediction for 5010 samples... 0.31251997 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 349 [0/25046 (0%)] Loss: 0.139653 Train epoch: 349 [328580/25046 (41%)] Loss: 0.131090 Train epoch: 349 [654960/25046 (82%)] Loss: 0.157545 Make prediction for 5010 samples... 0.33714986 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 350 [0/25046 (0%)] Loss: 0.159453 Train epoch: 350 [329160/25046 (41%)] Loss: 0.145450 Train epoch: 350 [670640/25046 (82%)] Loss: 0.158089 Make prediction for 5010 samples... 0.3090622 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 351 [0/25046 (0%)] Loss: 0.153467 Train epoch: 351 [328680/25046 (41%)] Loss: 0.141651 Train epoch: 351 [655040/25046 (82%)] Loss: 0.106261 Make prediction for 5010 samples... 0.3164035 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 352 [0/25046 (0%)] Loss: 0.127750 Train epoch: 352 [327820/25046 (41%)] Loss: 0.149701 Train epoch: 352 [649960/25046 (82%)] Loss: 0.152846 Make prediction for 5010 samples... 0.29095787 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 353 [0/25046 (0%)] Loss: 0.135581 Train epoch: 353 [327280/25046 (41%)] Loss: 0.155230 Train epoch: 353 [655680/25046 (82%)] Loss: 0.133536 Make prediction for 5010 samples... 0.30205017 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 354 [0/25046 (0%)] Loss: 0.208977 Train epoch: 354 [329520/25046 (41%)] Loss: 0.155268 Train epoch: 354 [657440/25046 (82%)] Loss: 0.137819 Make prediction for 5010 samples... 0.29790834 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 355 [0/25046 (0%)] Loss: 0.123243 Train epoch: 355 [327860/25046 (41%)] Loss: 0.160794 Train epoch: 355 [660680/25046 (82%)] Loss: 0.164953 Make prediction for 5010 samples... 0.3445717 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 356 [0/25046 (0%)] Loss: 0.120779 Train epoch: 356 [326800/25046 (41%)] Loss: 0.211190 Train epoch: 356 [659760/25046 (82%)] Loss: 0.152133 Make prediction for 5010 samples... 0.29876208 No improvement since epoch 342 ; best_mse,best_ci: 0.28569812 0.8697755226219489 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 357 [0/25046 (0%)] Loss: 0.158623 Train epoch: 357 [327700/25046 (41%)] Loss: 0.161643 Train epoch: 357 [662960/25046 (82%)] Loss: 0.129297 Make prediction for 5010 samples... rmse improved at epoch 357 ; best_mse,best_ci: 0.2841912 0.8712522205220575 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 358 [0/25046 (0%)] Loss: 0.101626 Train epoch: 358 [324540/25046 (41%)] Loss: 0.135548 Train epoch: 358 [675560/25046 (82%)] Loss: 0.148111 Make prediction for 5010 samples... 0.3084311 No improvement since epoch 357 ; best_mse,best_ci: 0.2841912 0.8712522205220575 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 359 [0/25046 (0%)] Loss: 0.189118 Train epoch: 359 [329060/25046 (41%)] Loss: 0.127320 Train epoch: 359 [664600/25046 (82%)] Loss: 0.143511 Make prediction for 5010 samples... 0.3020101 No improvement since epoch 357 ; best_mse,best_ci: 0.2841912 0.8712522205220575 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 360 [0/25046 (0%)] Loss: 0.136905 Train epoch: 360 [328500/25046 (41%)] Loss: 0.146905 Train epoch: 360 [660120/25046 (82%)] Loss: 0.153013 Make prediction for 5010 samples... 0.28884345 No improvement since epoch 357 ; best_mse,best_ci: 0.2841912 0.8712522205220575 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 361 [0/25046 (0%)] Loss: 0.130668 Train epoch: 361 [327160/25046 (41%)] Loss: 0.128011 Train epoch: 361 [662640/25046 (82%)] Loss: 0.183012 Make prediction for 5010 samples... 0.29098794 No improvement since epoch 357 ; best_mse,best_ci: 0.2841912 0.8712522205220575 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 362 [0/25046 (0%)] Loss: 0.131821 Train epoch: 362 [327840/25046 (41%)] Loss: 0.189877 Train epoch: 362 [650280/25046 (82%)] Loss: 0.129736 Make prediction for 5010 samples... 0.29202992 No improvement since epoch 357 ; best_mse,best_ci: 0.2841912 0.8712522205220575 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 363 [0/25046 (0%)] Loss: 0.103469 Train epoch: 363 [327500/25046 (41%)] Loss: 0.136400 Train epoch: 363 [657600/25046 (82%)] Loss: 0.140927 Make prediction for 5010 samples... rmse improved at epoch 363 ; best_mse,best_ci: 0.28183204 0.8728570306181257 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 364 [0/25046 (0%)] Loss: 0.146880 Train epoch: 364 [322980/25046 (41%)] Loss: 0.137990 Train epoch: 364 [667080/25046 (82%)] Loss: 0.179150 Make prediction for 5010 samples... 0.28531316 No improvement since epoch 363 ; best_mse,best_ci: 0.28183204 0.8728570306181257 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 365 [0/25046 (0%)] Loss: 0.137778 Train epoch: 365 [328600/25046 (41%)] Loss: 0.130398 Train epoch: 365 [665960/25046 (82%)] Loss: 0.171778 Make prediction for 5010 samples... 0.29720163 No improvement since epoch 363 ; best_mse,best_ci: 0.28183204 0.8728570306181257 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 366 [0/25046 (0%)] Loss: 0.131456 Train epoch: 366 [331400/25046 (41%)] Loss: 0.163574 Train epoch: 366 [655560/25046 (82%)] Loss: 0.188040 Make prediction for 5010 samples... 0.33592445 No improvement since epoch 363 ; best_mse,best_ci: 0.28183204 0.8728570306181257 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 367 [0/25046 (0%)] Loss: 0.183187 Train epoch: 367 [328960/25046 (41%)] Loss: 0.130960 Train epoch: 367 [665600/25046 (82%)] Loss: 0.152219 Make prediction for 5010 samples... 0.29425097 No improvement since epoch 363 ; best_mse,best_ci: 0.28183204 0.8728570306181257 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 368 [0/25046 (0%)] Loss: 0.144769 Train epoch: 368 [333520/25046 (41%)] Loss: 0.151347 Train epoch: 368 [656640/25046 (82%)] Loss: 0.167894 Make prediction for 5010 samples... 0.31131658 No improvement since epoch 363 ; best_mse,best_ci: 0.28183204 0.8728570306181257 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 369 [0/25046 (0%)] Loss: 0.123476 Train epoch: 369 [331320/25046 (41%)] Loss: 0.147339 Train epoch: 369 [665440/25046 (82%)] Loss: 0.115212 Make prediction for 5010 samples... 0.2831643 No improvement since epoch 363 ; best_mse,best_ci: 0.28183204 0.8728570306181257 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 370 [0/25046 (0%)] Loss: 0.128321 Train epoch: 370 [327680/25046 (41%)] Loss: 0.141230 Train epoch: 370 [651360/25046 (82%)] Loss: 0.151046 Make prediction for 5010 samples... 0.35269874 No improvement since epoch 363 ; best_mse,best_ci: 0.28183204 0.8728570306181257 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 371 [0/25046 (0%)] Loss: 0.197546 Train epoch: 371 [335320/25046 (41%)] Loss: 0.142695 Train epoch: 371 [660120/25046 (82%)] Loss: 0.166724 Make prediction for 5010 samples... rmse improved at epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 372 [0/25046 (0%)] Loss: 0.116846 Train epoch: 372 [333040/25046 (41%)] Loss: 0.147477 Train epoch: 372 [649200/25046 (82%)] Loss: 0.162262 Make prediction for 5010 samples... 0.32409835 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 373 [0/25046 (0%)] Loss: 0.162421 Train epoch: 373 [330120/25046 (41%)] Loss: 0.170807 Train epoch: 373 [645200/25046 (82%)] Loss: 0.146004 Make prediction for 5010 samples... 0.30228025 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 374 [0/25046 (0%)] Loss: 0.120571 Train epoch: 374 [329140/25046 (41%)] Loss: 0.133594 Train epoch: 374 [650120/25046 (82%)] Loss: 0.142833 Make prediction for 5010 samples... 0.29638755 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 375 [0/25046 (0%)] Loss: 0.114797 Train epoch: 375 [329560/25046 (41%)] Loss: 0.140677 Train epoch: 375 [653240/25046 (82%)] Loss: 0.177394 Make prediction for 5010 samples... 0.31633887 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 376 [0/25046 (0%)] Loss: 0.129297 Train epoch: 376 [329560/25046 (41%)] Loss: 0.140315 Train epoch: 376 [649800/25046 (82%)] Loss: 0.171669 Make prediction for 5010 samples... 0.28221235 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 377 [0/25046 (0%)] Loss: 0.141053 Train epoch: 377 [326940/25046 (41%)] Loss: 0.158173 Train epoch: 377 [661720/25046 (82%)] Loss: 0.155596 Make prediction for 5010 samples... 0.34144524 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 378 [0/25046 (0%)] Loss: 0.164924 Train epoch: 378 [324780/25046 (41%)] Loss: 0.168474 Train epoch: 378 [655400/25046 (82%)] Loss: 0.139868 Make prediction for 5010 samples... 0.30735266 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 379 [0/25046 (0%)] Loss: 0.150960 Train epoch: 379 [332600/25046 (41%)] Loss: 0.215046 Train epoch: 379 [657560/25046 (82%)] Loss: 0.141181 Make prediction for 5010 samples... 0.29475275 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 380 [0/25046 (0%)] Loss: 0.130322 Train epoch: 380 [329060/25046 (41%)] Loss: 0.154902 Train epoch: 380 [659760/25046 (82%)] Loss: 0.166843 Make prediction for 5010 samples... 0.30023694 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 381 [0/25046 (0%)] Loss: 0.160424 Train epoch: 381 [331600/25046 (41%)] Loss: 0.159157 Train epoch: 381 [651400/25046 (82%)] Loss: 0.171444 Make prediction for 5010 samples... 0.29374412 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 382 [0/25046 (0%)] Loss: 0.118174 Train epoch: 382 [326900/25046 (41%)] Loss: 0.162493 Train epoch: 382 [650680/25046 (82%)] Loss: 0.148645 Make prediction for 5010 samples... 0.2803928 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 383 [0/25046 (0%)] Loss: 0.132555 Train epoch: 383 [324440/25046 (41%)] Loss: 0.142978 Train epoch: 383 [665480/25046 (82%)] Loss: 0.143551 Make prediction for 5010 samples... 0.30512828 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 384 [0/25046 (0%)] Loss: 0.131406 Train epoch: 384 [326740/25046 (41%)] Loss: 0.165232 Train epoch: 384 [662320/25046 (82%)] Loss: 0.161102 Make prediction for 5010 samples... 0.29892012 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 385 [0/25046 (0%)] Loss: 0.127188 Train epoch: 385 [327060/25046 (41%)] Loss: 0.135566 Train epoch: 385 [659800/25046 (82%)] Loss: 0.154608 Make prediction for 5010 samples... 0.3039597 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 386 [0/25046 (0%)] Loss: 0.142206 Train epoch: 386 [328460/25046 (41%)] Loss: 0.123516 Train epoch: 386 [645840/25046 (82%)] Loss: 0.149187 Make prediction for 5010 samples... 0.35630447 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 387 [0/25046 (0%)] Loss: 0.139942 Train epoch: 387 [327180/25046 (41%)] Loss: 0.147363 Train epoch: 387 [652760/25046 (82%)] Loss: 0.160501 Make prediction for 5010 samples... 0.29090944 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 388 [0/25046 (0%)] Loss: 0.135588 Train epoch: 388 [333200/25046 (41%)] Loss: 0.184626 Train epoch: 388 [653320/25046 (82%)] Loss: 0.151371 Make prediction for 5010 samples... 0.31709307 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 389 [0/25046 (0%)] Loss: 0.146858 Train epoch: 389 [327340/25046 (41%)] Loss: 0.133378 Train epoch: 389 [648720/25046 (82%)] Loss: 0.168606 Make prediction for 5010 samples... 0.2792665 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 390 [0/25046 (0%)] Loss: 0.124446 Train epoch: 390 [324320/25046 (41%)] Loss: 0.111601 Train epoch: 390 [663680/25046 (82%)] Loss: 0.111053 Make prediction for 5010 samples... 0.28441462 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 391 [0/25046 (0%)] Loss: 0.141405 Train epoch: 391 [333200/25046 (41%)] Loss: 0.127685 Train epoch: 391 [652200/25046 (82%)] Loss: 0.153704 Make prediction for 5010 samples... 0.29267988 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 392 [0/25046 (0%)] Loss: 0.122576 Train epoch: 392 [335040/25046 (41%)] Loss: 0.126302 Train epoch: 392 [658840/25046 (82%)] Loss: 0.138309 Make prediction for 5010 samples... 0.296801 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 393 [0/25046 (0%)] Loss: 0.117678 Train epoch: 393 [328520/25046 (41%)] Loss: 0.176464 Train epoch: 393 [664800/25046 (82%)] Loss: 0.139516 Make prediction for 5010 samples... 0.2794223 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 394 [0/25046 (0%)] Loss: 0.132048 Train epoch: 394 [326900/25046 (41%)] Loss: 0.150881 Train epoch: 394 [655120/25046 (82%)] Loss: 0.112531 Make prediction for 5010 samples... 0.29336095 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 395 [0/25046 (0%)] Loss: 0.125813 Train epoch: 395 [331060/25046 (41%)] Loss: 0.126136 Train epoch: 395 [652000/25046 (82%)] Loss: 0.152045 Make prediction for 5010 samples... 0.28152707 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 396 [0/25046 (0%)] Loss: 0.105721 Train epoch: 396 [328360/25046 (41%)] Loss: 0.164892 Train epoch: 396 [652040/25046 (82%)] Loss: 0.158187 Make prediction for 5010 samples... 0.2910873 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 397 [0/25046 (0%)] Loss: 0.122621 Train epoch: 397 [330260/25046 (41%)] Loss: 0.145201 Train epoch: 397 [660760/25046 (82%)] Loss: 0.109983 Make prediction for 5010 samples... 0.2942145 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 398 [0/25046 (0%)] Loss: 0.161788 Train epoch: 398 [323080/25046 (41%)] Loss: 0.130643 Train epoch: 398 [666200/25046 (82%)] Loss: 0.168522 Make prediction for 5010 samples... 0.2905888 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 399 [0/25046 (0%)] Loss: 0.165788 Train epoch: 399 [326260/25046 (41%)] Loss: 0.161345 Train epoch: 399 [663160/25046 (82%)] Loss: 0.140384 Make prediction for 5010 samples... 0.29093552 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 400 [0/25046 (0%)] Loss: 0.121667 Train epoch: 400 [333080/25046 (41%)] Loss: 0.141020 Train epoch: 400 [651280/25046 (82%)] Loss: 0.173414 Make prediction for 5010 samples... 0.2782595 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 401 [0/25046 (0%)] Loss: 0.140603 Train epoch: 401 [327460/25046 (41%)] Loss: 0.123340 Train epoch: 401 [643960/25046 (82%)] Loss: 0.152850 Make prediction for 5010 samples... 0.28585014 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 402 [0/25046 (0%)] Loss: 0.126017 Train epoch: 402 [325140/25046 (41%)] Loss: 0.122152 Train epoch: 402 [650480/25046 (82%)] Loss: 0.134807 Make prediction for 5010 samples... 0.2795705 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 403 [0/25046 (0%)] Loss: 0.156113 Train epoch: 403 [326300/25046 (41%)] Loss: 0.115386 Train epoch: 403 [656880/25046 (82%)] Loss: 0.147612 Make prediction for 5010 samples... 0.301357 No improvement since epoch 371 ; best_mse,best_ci: 0.27772015 0.8622141463125512 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 404 [0/25046 (0%)] Loss: 0.126060 Train epoch: 404 [325680/25046 (41%)] Loss: 0.101416 Train epoch: 404 [650240/25046 (82%)] Loss: 0.180318 Make prediction for 5010 samples... rmse improved at epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 405 [0/25046 (0%)] Loss: 0.126890 Train epoch: 405 [329540/25046 (41%)] Loss: 0.149389 Train epoch: 405 [651280/25046 (82%)] Loss: 0.122847 Make prediction for 5010 samples... 0.2869088 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 406 [0/25046 (0%)] Loss: 0.155259 Train epoch: 406 [332400/25046 (41%)] Loss: 0.156830 Train epoch: 406 [664360/25046 (82%)] Loss: 0.152301 Make prediction for 5010 samples... 0.2971915 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 407 [0/25046 (0%)] Loss: 0.115411 Train epoch: 407 [329500/25046 (41%)] Loss: 0.148303 Train epoch: 407 [662440/25046 (82%)] Loss: 0.156671 Make prediction for 5010 samples... 0.2811731 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 408 [0/25046 (0%)] Loss: 0.135340 Train epoch: 408 [330480/25046 (41%)] Loss: 0.147232 Train epoch: 408 [660720/25046 (82%)] Loss: 0.151279 Make prediction for 5010 samples... 0.3088123 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 409 [0/25046 (0%)] Loss: 0.141634 Train epoch: 409 [332340/25046 (41%)] Loss: 0.115815 Train epoch: 409 [653480/25046 (82%)] Loss: 0.144993 Make prediction for 5010 samples... 0.2842038 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 410 [0/25046 (0%)] Loss: 0.113096 Train epoch: 410 [326660/25046 (41%)] Loss: 0.137496 Train epoch: 410 [653800/25046 (82%)] Loss: 0.142890 Make prediction for 5010 samples... 0.28405526 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 411 [0/25046 (0%)] Loss: 0.133551 Train epoch: 411 [328220/25046 (41%)] Loss: 0.139496 Train epoch: 411 [653520/25046 (82%)] Loss: 0.164660 Make prediction for 5010 samples... 0.2912471 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 412 [0/25046 (0%)] Loss: 0.137842 Train epoch: 412 [328040/25046 (41%)] Loss: 0.159012 Train epoch: 412 [652960/25046 (82%)] Loss: 0.161520 Make prediction for 5010 samples... 0.28121316 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 413 [0/25046 (0%)] Loss: 0.143544 Train epoch: 413 [330420/25046 (41%)] Loss: 0.159556 Train epoch: 413 [651640/25046 (82%)] Loss: 0.175147 Make prediction for 5010 samples... 0.29516718 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 414 [0/25046 (0%)] Loss: 0.119986 Train epoch: 414 [329820/25046 (41%)] Loss: 0.115070 Train epoch: 414 [658840/25046 (82%)] Loss: 0.111854 Make prediction for 5010 samples... 0.2866708 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 415 [0/25046 (0%)] Loss: 0.114007 Train epoch: 415 [326360/25046 (41%)] Loss: 0.126608 Train epoch: 415 [661840/25046 (82%)] Loss: 0.123197 Make prediction for 5010 samples... 0.28502366 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 416 [0/25046 (0%)] Loss: 0.102696 Train epoch: 416 [329440/25046 (41%)] Loss: 0.125070 Train epoch: 416 [655960/25046 (82%)] Loss: 0.125020 Make prediction for 5010 samples... 0.29804415 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 417 [0/25046 (0%)] Loss: 0.125892 Train epoch: 417 [332620/25046 (41%)] Loss: 0.130360 Train epoch: 417 [662680/25046 (82%)] Loss: 0.171995 Make prediction for 5010 samples... 0.28681108 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 418 [0/25046 (0%)] Loss: 0.167778 Train epoch: 418 [331280/25046 (41%)] Loss: 0.099277 Train epoch: 418 [654040/25046 (82%)] Loss: 0.127252 Make prediction for 5010 samples... 0.28121683 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 419 [0/25046 (0%)] Loss: 0.111948 Train epoch: 419 [326560/25046 (41%)] Loss: 0.097562 Train epoch: 419 [661120/25046 (82%)] Loss: 0.137330 Make prediction for 5010 samples... 0.27987617 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 420 [0/25046 (0%)] Loss: 0.134846 Train epoch: 420 [327920/25046 (41%)] Loss: 0.114120 Train epoch: 420 [653040/25046 (82%)] Loss: 0.138007 Make prediction for 5010 samples... 0.27965868 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 421 [0/25046 (0%)] Loss: 0.160635 Train epoch: 421 [329460/25046 (41%)] Loss: 0.137024 Train epoch: 421 [658400/25046 (82%)] Loss: 0.122456 Make prediction for 5010 samples... 0.3102731 No improvement since epoch 404 ; best_mse,best_ci: 0.27133942 0.8675831423647352 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 422 [0/25046 (0%)] Loss: 0.139081 Train epoch: 422 [329120/25046 (41%)] Loss: 0.121988 Train epoch: 422 [659240/25046 (82%)] Loss: 0.127770 Make prediction for 5010 samples... rmse improved at epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 423 [0/25046 (0%)] Loss: 0.124997 Train epoch: 423 [325360/25046 (41%)] Loss: 0.108778 Train epoch: 423 [655760/25046 (82%)] Loss: 0.112165 Make prediction for 5010 samples... 0.28581017 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 424 [0/25046 (0%)] Loss: 0.151478 Train epoch: 424 [333040/25046 (41%)] Loss: 0.139622 Train epoch: 424 [664120/25046 (82%)] Loss: 0.132103 Make prediction for 5010 samples... 0.31941113 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 425 [0/25046 (0%)] Loss: 0.148374 Train epoch: 425 [325760/25046 (41%)] Loss: 0.190381 Train epoch: 425 [656560/25046 (82%)] Loss: 0.140744 Make prediction for 5010 samples... 0.2843372 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 426 [0/25046 (0%)] Loss: 0.132927 Train epoch: 426 [329840/25046 (41%)] Loss: 0.138916 Train epoch: 426 [662560/25046 (82%)] Loss: 0.156172 Make prediction for 5010 samples... 0.2974034 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 427 [0/25046 (0%)] Loss: 0.154675 Train epoch: 427 [328740/25046 (41%)] Loss: 0.141938 Train epoch: 427 [662360/25046 (82%)] Loss: 0.139422 Make prediction for 5010 samples... 0.29168075 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 428 [0/25046 (0%)] Loss: 0.115658 Train epoch: 428 [323300/25046 (41%)] Loss: 0.174646 Train epoch: 428 [665200/25046 (82%)] Loss: 0.179571 Make prediction for 5010 samples... 0.27275857 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 429 [0/25046 (0%)] Loss: 0.133000 Train epoch: 429 [331580/25046 (41%)] Loss: 0.143364 Train epoch: 429 [657560/25046 (82%)] Loss: 0.152028 Make prediction for 5010 samples... 0.30426198 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 430 [0/25046 (0%)] Loss: 0.141270 Train epoch: 430 [324280/25046 (41%)] Loss: 0.105616 Train epoch: 430 [647520/25046 (82%)] Loss: 0.138185 Make prediction for 5010 samples... 0.2868193 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 431 [0/25046 (0%)] Loss: 0.129472 Train epoch: 431 [325340/25046 (41%)] Loss: 0.166067 Train epoch: 431 [672800/25046 (82%)] Loss: 0.197920 Make prediction for 5010 samples... 0.30386424 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 432 [0/25046 (0%)] Loss: 0.166342 Train epoch: 432 [325680/25046 (41%)] Loss: 0.181010 Train epoch: 432 [659920/25046 (82%)] Loss: 0.135742 Make prediction for 5010 samples... 0.29149187 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 433 [0/25046 (0%)] Loss: 0.171718 Train epoch: 433 [332680/25046 (41%)] Loss: 0.211116 Train epoch: 433 [655320/25046 (82%)] Loss: 0.159662 Make prediction for 5010 samples... 0.29250893 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 434 [0/25046 (0%)] Loss: 0.157584 Train epoch: 434 [328240/25046 (41%)] Loss: 0.160010 Train epoch: 434 [655360/25046 (82%)] Loss: 0.144602 Make prediction for 5010 samples... 0.3019808 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 435 [0/25046 (0%)] Loss: 0.104233 Train epoch: 435 [325800/25046 (41%)] Loss: 0.119021 Train epoch: 435 [657840/25046 (82%)] Loss: 0.143526 Make prediction for 5010 samples... 0.29760027 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 436 [0/25046 (0%)] Loss: 0.118276 Train epoch: 436 [326900/25046 (41%)] Loss: 0.114817 Train epoch: 436 [652600/25046 (82%)] Loss: 0.139770 Make prediction for 5010 samples... 0.35329065 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 437 [0/25046 (0%)] Loss: 0.199064 Train epoch: 437 [325480/25046 (41%)] Loss: 0.180212 Train epoch: 437 [648080/25046 (82%)] Loss: 0.154305 Make prediction for 5010 samples... 0.28573778 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 438 [0/25046 (0%)] Loss: 0.141484 Train epoch: 438 [327240/25046 (41%)] Loss: 0.119984 Train epoch: 438 [652760/25046 (82%)] Loss: 0.156815 Make prediction for 5010 samples... 0.2823363 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 439 [0/25046 (0%)] Loss: 0.131043 Train epoch: 439 [329300/25046 (41%)] Loss: 0.119717 Train epoch: 439 [650720/25046 (82%)] Loss: 0.134663 Make prediction for 5010 samples... 0.30448523 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 440 [0/25046 (0%)] Loss: 0.127687 Train epoch: 440 [327820/25046 (41%)] Loss: 0.154996 Train epoch: 440 [658640/25046 (82%)] Loss: 0.140881 Make prediction for 5010 samples... 0.2982109 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 441 [0/25046 (0%)] Loss: 0.109511 Train epoch: 441 [324440/25046 (41%)] Loss: 0.123655 Train epoch: 441 [664080/25046 (82%)] Loss: 0.124141 Make prediction for 5010 samples... 0.27714074 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 442 [0/25046 (0%)] Loss: 0.118799 Train epoch: 442 [323440/25046 (41%)] Loss: 0.162342 Train epoch: 442 [664480/25046 (82%)] Loss: 0.144431 Make prediction for 5010 samples... 0.28597987 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 443 [0/25046 (0%)] Loss: 0.119394 Train epoch: 443 [336520/25046 (41%)] Loss: 0.141866 Train epoch: 443 [661120/25046 (82%)] Loss: 0.116147 Make prediction for 5010 samples... 0.28672475 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 444 [0/25046 (0%)] Loss: 0.114332 Train epoch: 444 [332820/25046 (41%)] Loss: 0.180729 Train epoch: 444 [663320/25046 (82%)] Loss: 0.148058 Make prediction for 5010 samples... 0.3065557 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 445 [0/25046 (0%)] Loss: 0.172010 Train epoch: 445 [326440/25046 (41%)] Loss: 0.132613 Train epoch: 445 [655680/25046 (82%)] Loss: 0.131356 Make prediction for 5010 samples... 0.28281063 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 446 [0/25046 (0%)] Loss: 0.112524 Train epoch: 446 [322680/25046 (41%)] Loss: 0.118608 Train epoch: 446 [653840/25046 (82%)] Loss: 0.119617 Make prediction for 5010 samples... 0.2914388 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 447 [0/25046 (0%)] Loss: 0.125034 Train epoch: 447 [331860/25046 (41%)] Loss: 0.153614 Train epoch: 447 [660120/25046 (82%)] Loss: 0.139307 Make prediction for 5010 samples... 0.27649266 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 448 [0/25046 (0%)] Loss: 0.140836 Train epoch: 448 [324580/25046 (41%)] Loss: 0.126453 Train epoch: 448 [663400/25046 (82%)] Loss: 0.170599 Make prediction for 5010 samples... 0.30470386 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 449 [0/25046 (0%)] Loss: 0.118593 Train epoch: 449 [327760/25046 (41%)] Loss: 0.173051 Train epoch: 449 [659200/25046 (82%)] Loss: 0.171014 Make prediction for 5010 samples... 0.34023008 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 450 [0/25046 (0%)] Loss: 0.149259 Train epoch: 450 [327560/25046 (41%)] Loss: 0.150523 Train epoch: 450 [656160/25046 (82%)] Loss: 0.139309 Make prediction for 5010 samples... 0.28086615 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 451 [0/25046 (0%)] Loss: 0.146665 Train epoch: 451 [323700/25046 (41%)] Loss: 0.173211 Train epoch: 451 [660040/25046 (82%)] Loss: 0.159537 Make prediction for 5010 samples... 0.28143036 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 452 [0/25046 (0%)] Loss: 0.097079 Train epoch: 452 [326360/25046 (41%)] Loss: 0.129411 Train epoch: 452 [656600/25046 (82%)] Loss: 0.129500 Make prediction for 5010 samples... 0.33391505 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 453 [0/25046 (0%)] Loss: 0.153718 Train epoch: 453 [325620/25046 (41%)] Loss: 0.169305 Train epoch: 453 [651040/25046 (82%)] Loss: 0.122961 Make prediction for 5010 samples... 0.28480515 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 454 [0/25046 (0%)] Loss: 0.124761 Train epoch: 454 [326800/25046 (41%)] Loss: 0.131910 Train epoch: 454 [649360/25046 (82%)] Loss: 0.160824 Make prediction for 5010 samples... 0.3040431 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 455 [0/25046 (0%)] Loss: 0.126533 Train epoch: 455 [328540/25046 (41%)] Loss: 0.185368 Train epoch: 455 [644640/25046 (82%)] Loss: 0.133778 Make prediction for 5010 samples... 0.30047384 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 456 [0/25046 (0%)] Loss: 0.139917 Train epoch: 456 [332840/25046 (41%)] Loss: 0.122351 Train epoch: 456 [661520/25046 (82%)] Loss: 0.153430 Make prediction for 5010 samples... 0.31447238 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 457 [0/25046 (0%)] Loss: 0.097437 Train epoch: 457 [325860/25046 (41%)] Loss: 0.129819 Train epoch: 457 [660160/25046 (82%)] Loss: 0.140647 Make prediction for 5010 samples... 0.2999988 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 458 [0/25046 (0%)] Loss: 0.135147 Train epoch: 458 [326060/25046 (41%)] Loss: 0.126514 Train epoch: 458 [650240/25046 (82%)] Loss: 0.132351 Make prediction for 5010 samples... 0.33980837 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 459 [0/25046 (0%)] Loss: 0.134139 Train epoch: 459 [325160/25046 (41%)] Loss: 0.103896 Train epoch: 459 [650240/25046 (82%)] Loss: 0.171502 Make prediction for 5010 samples... 0.27314916 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 460 [0/25046 (0%)] Loss: 0.111686 Train epoch: 460 [329720/25046 (41%)] Loss: 0.112518 Train epoch: 460 [654280/25046 (82%)] Loss: 0.137457 Make prediction for 5010 samples... 0.33174455 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 461 [0/25046 (0%)] Loss: 0.123998 Train epoch: 461 [332080/25046 (41%)] Loss: 0.093374 Train epoch: 461 [664640/25046 (82%)] Loss: 0.140560 Make prediction for 5010 samples... 0.32186276 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 462 [0/25046 (0%)] Loss: 0.156476 Train epoch: 462 [324500/25046 (41%)] Loss: 0.147640 Train epoch: 462 [660440/25046 (82%)] Loss: 0.156019 Make prediction for 5010 samples... 0.2957528 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 463 [0/25046 (0%)] Loss: 0.117600 Train epoch: 463 [328300/25046 (41%)] Loss: 0.108455 Train epoch: 463 [660000/25046 (82%)] Loss: 0.129021 Make prediction for 5010 samples... 0.2823112 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 464 [0/25046 (0%)] Loss: 0.132637 Train epoch: 464 [323800/25046 (41%)] Loss: 0.099208 Train epoch: 464 [646240/25046 (82%)] Loss: 0.113318 Make prediction for 5010 samples... 0.2863019 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 465 [0/25046 (0%)] Loss: 0.155913 Train epoch: 465 [335560/25046 (41%)] Loss: 0.129670 Train epoch: 465 [658880/25046 (82%)] Loss: 0.147670 Make prediction for 5010 samples... 0.32652164 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 466 [0/25046 (0%)] Loss: 0.122289 Train epoch: 466 [329160/25046 (41%)] Loss: 0.103874 Train epoch: 466 [652560/25046 (82%)] Loss: 0.177977 Make prediction for 5010 samples... 0.30121976 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 467 [0/25046 (0%)] Loss: 0.181341 Train epoch: 467 [331360/25046 (41%)] Loss: 0.192302 Train epoch: 467 [665800/25046 (82%)] Loss: 0.127774 Make prediction for 5010 samples... 0.29538655 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 468 [0/25046 (0%)] Loss: 0.141873 Train epoch: 468 [325580/25046 (41%)] Loss: 0.159999 Train epoch: 468 [655640/25046 (82%)] Loss: 0.132182 Make prediction for 5010 samples... 0.27974963 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 469 [0/25046 (0%)] Loss: 0.087762 Train epoch: 469 [325380/25046 (41%)] Loss: 0.142108 Train epoch: 469 [657120/25046 (82%)] Loss: 0.159821 Make prediction for 5010 samples... 0.28982908 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 470 [0/25046 (0%)] Loss: 0.110751 Train epoch: 470 [325240/25046 (41%)] Loss: 0.108528 Train epoch: 470 [654560/25046 (82%)] Loss: 0.096050 Make prediction for 5010 samples... 0.29111484 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 471 [0/25046 (0%)] Loss: 0.159677 Train epoch: 471 [329920/25046 (41%)] Loss: 0.157057 Train epoch: 471 [655800/25046 (82%)] Loss: 0.116011 Make prediction for 5010 samples... 0.2865604 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 472 [0/25046 (0%)] Loss: 0.105584 Train epoch: 472 [326600/25046 (41%)] Loss: 0.122602 Train epoch: 472 [658720/25046 (82%)] Loss: 0.133501 Make prediction for 5010 samples... 0.2910948 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 473 [0/25046 (0%)] Loss: 0.151742 Train epoch: 473 [330060/25046 (41%)] Loss: 0.125376 Train epoch: 473 [652920/25046 (82%)] Loss: 0.141199 Make prediction for 5010 samples... 0.288814 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 474 [0/25046 (0%)] Loss: 0.120829 Train epoch: 474 [328340/25046 (41%)] Loss: 0.126356 Train epoch: 474 [655760/25046 (82%)] Loss: 0.108748 Make prediction for 5010 samples... 0.295971 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 475 [0/25046 (0%)] Loss: 0.120250 Train epoch: 475 [325940/25046 (41%)] Loss: 0.166723 Train epoch: 475 [660960/25046 (82%)] Loss: 0.110673 Make prediction for 5010 samples... 0.2827335 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 476 [0/25046 (0%)] Loss: 0.111983 Train epoch: 476 [325480/25046 (41%)] Loss: 0.153630 Train epoch: 476 [653040/25046 (82%)] Loss: 0.139049 Make prediction for 5010 samples... 0.28646412 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 477 [0/25046 (0%)] Loss: 0.129041 Train epoch: 477 [335780/25046 (41%)] Loss: 0.117233 Train epoch: 477 [654840/25046 (82%)] Loss: 0.150311 Make prediction for 5010 samples... 0.306204 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 478 [0/25046 (0%)] Loss: 0.194186 Train epoch: 478 [331540/25046 (41%)] Loss: 0.129316 Train epoch: 478 [659360/25046 (82%)] Loss: 0.203724 Make prediction for 5010 samples... 0.28683218 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 479 [0/25046 (0%)] Loss: 0.129337 Train epoch: 479 [323360/25046 (41%)] Loss: 0.140429 Train epoch: 479 [657800/25046 (82%)] Loss: 0.158286 Make prediction for 5010 samples... 0.3108304 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 480 [0/25046 (0%)] Loss: 0.185685 Train epoch: 480 [324580/25046 (41%)] Loss: 0.146381 Train epoch: 480 [648840/25046 (82%)] Loss: 0.142177 Make prediction for 5010 samples... 0.28248557 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 481 [0/25046 (0%)] Loss: 0.135615 Train epoch: 481 [324920/25046 (41%)] Loss: 0.128002 Train epoch: 481 [658160/25046 (82%)] Loss: 0.128719 Make prediction for 5010 samples... 0.28767624 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 482 [0/25046 (0%)] Loss: 0.134017 Train epoch: 482 [330880/25046 (41%)] Loss: 0.146931 Train epoch: 482 [659640/25046 (82%)] Loss: 0.119192 Make prediction for 5010 samples... 0.29462594 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 483 [0/25046 (0%)] Loss: 0.109234 Train epoch: 483 [326100/25046 (41%)] Loss: 0.106956 Train epoch: 483 [658600/25046 (82%)] Loss: 0.146085 Make prediction for 5010 samples... 0.34879422 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 484 [0/25046 (0%)] Loss: 0.188473 Train epoch: 484 [331420/25046 (41%)] Loss: 0.099593 Train epoch: 484 [663920/25046 (82%)] Loss: 0.162060 Make prediction for 5010 samples... 0.2896982 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 485 [0/25046 (0%)] Loss: 0.162145 Train epoch: 485 [325520/25046 (41%)] Loss: 0.141683 Train epoch: 485 [645600/25046 (82%)] Loss: 0.124179 Make prediction for 5010 samples... 0.33049688 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 486 [0/25046 (0%)] Loss: 0.128119 Train epoch: 486 [330160/25046 (41%)] Loss: 0.104213 Train epoch: 486 [655320/25046 (82%)] Loss: 0.122775 Make prediction for 5010 samples... 0.28313443 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 487 [0/25046 (0%)] Loss: 0.131016 Train epoch: 487 [326160/25046 (41%)] Loss: 0.103020 Train epoch: 487 [661720/25046 (82%)] Loss: 0.132027 Make prediction for 5010 samples... 0.28745627 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 488 [0/25046 (0%)] Loss: 0.108244 Train epoch: 488 [327200/25046 (41%)] Loss: 0.113293 Train epoch: 488 [661480/25046 (82%)] Loss: 0.104972 Make prediction for 5010 samples... 0.2801843 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 489 [0/25046 (0%)] Loss: 0.135865 Train epoch: 489 [327620/25046 (41%)] Loss: 0.162441 Train epoch: 489 [661960/25046 (82%)] Loss: 0.115883 Make prediction for 5010 samples... 0.28180218 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 490 [0/25046 (0%)] Loss: 0.105259 Train epoch: 490 [328280/25046 (41%)] Loss: 0.108192 Train epoch: 490 [659080/25046 (82%)] Loss: 0.125962 Make prediction for 5010 samples... 0.28509492 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 491 [0/25046 (0%)] Loss: 0.128038 Train epoch: 491 [329520/25046 (41%)] Loss: 0.111157 Train epoch: 491 [661280/25046 (82%)] Loss: 0.131530 Make prediction for 5010 samples... 0.28895572 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 492 [0/25046 (0%)] Loss: 0.147301 Train epoch: 492 [329940/25046 (41%)] Loss: 0.129934 Train epoch: 492 [654960/25046 (82%)] Loss: 0.164562 Make prediction for 5010 samples... 0.32149655 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 493 [0/25046 (0%)] Loss: 0.120744 Train epoch: 493 [330420/25046 (41%)] Loss: 0.143720 Train epoch: 493 [672080/25046 (82%)] Loss: 0.110727 Make prediction for 5010 samples... 0.2878684 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 494 [0/25046 (0%)] Loss: 0.113711 Train epoch: 494 [329820/25046 (41%)] Loss: 0.125059 Train epoch: 494 [654600/25046 (82%)] Loss: 0.143802 Make prediction for 5010 samples... 0.30550894 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 495 [0/25046 (0%)] Loss: 0.124477 Train epoch: 495 [329920/25046 (41%)] Loss: 0.171378 Train epoch: 495 [650760/25046 (82%)] Loss: 0.124332 Make prediction for 5010 samples... 0.2780112 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 496 [0/25046 (0%)] Loss: 0.116821 Train epoch: 496 [329560/25046 (41%)] Loss: 0.130767 Train epoch: 496 [655000/25046 (82%)] Loss: 0.122910 Make prediction for 5010 samples... 0.2941771 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 497 [0/25046 (0%)] Loss: 0.114306 Train epoch: 497 [328720/25046 (41%)] Loss: 0.140432 Train epoch: 497 [659600/25046 (82%)] Loss: 0.135619 Make prediction for 5010 samples... 0.30702424 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 498 [0/25046 (0%)] Loss: 0.182163 Train epoch: 498 [330500/25046 (41%)] Loss: 0.121350 Train epoch: 498 [672000/25046 (82%)] Loss: 0.157864 Make prediction for 5010 samples... 0.28110477 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 499 [0/25046 (0%)] Loss: 0.135305 Train epoch: 499 [330140/25046 (41%)] Loss: 0.161119 Train epoch: 499 [658040/25046 (82%)] Loss: 0.109867 Make prediction for 5010 samples... 0.294337 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 500 [0/25046 (0%)] Loss: 0.164877 Train epoch: 500 [328300/25046 (41%)] Loss: 0.124737 Train epoch: 500 [658320/25046 (82%)] Loss: 0.123340 Make prediction for 5010 samples... 0.29821786 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 501 [0/25046 (0%)] Loss: 0.106486 Train epoch: 501 [325840/25046 (41%)] Loss: 0.093197 Train epoch: 501 [660720/25046 (82%)] Loss: 0.133166 Make prediction for 5010 samples... 0.28589088 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 502 [0/25046 (0%)] Loss: 0.110781 Train epoch: 502 [328140/25046 (41%)] Loss: 0.123514 Train epoch: 502 [656360/25046 (82%)] Loss: 0.161985 Make prediction for 5010 samples... 0.28390646 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 503 [0/25046 (0%)] Loss: 0.110509 Train epoch: 503 [327700/25046 (41%)] Loss: 0.156974 Train epoch: 503 [659280/25046 (82%)] Loss: 0.110093 Make prediction for 5010 samples... 0.30423245 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 504 [0/25046 (0%)] Loss: 0.115312 Train epoch: 504 [330120/25046 (41%)] Loss: 0.132969 Train epoch: 504 [658640/25046 (82%)] Loss: 0.106065 Make prediction for 5010 samples... 0.29154015 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 505 [0/25046 (0%)] Loss: 0.126160 Train epoch: 505 [333920/25046 (41%)] Loss: 0.117885 Train epoch: 505 [657960/25046 (82%)] Loss: 0.163557 Make prediction for 5010 samples... 0.28830543 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 506 [0/25046 (0%)] Loss: 0.101463 Train epoch: 506 [325320/25046 (41%)] Loss: 0.134971 Train epoch: 506 [658480/25046 (82%)] Loss: 0.124166 Make prediction for 5010 samples... 0.2810844 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 507 [0/25046 (0%)] Loss: 0.130953 Train epoch: 507 [329480/25046 (41%)] Loss: 0.110547 Train epoch: 507 [655720/25046 (82%)] Loss: 0.159314 Make prediction for 5010 samples... 0.2931657 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 508 [0/25046 (0%)] Loss: 0.146205 Train epoch: 508 [330920/25046 (41%)] Loss: 0.150731 Train epoch: 508 [656440/25046 (82%)] Loss: 0.136849 Make prediction for 5010 samples... 0.32621 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 509 [0/25046 (0%)] Loss: 0.142423 Train epoch: 509 [334880/25046 (41%)] Loss: 0.106113 Train epoch: 509 [651560/25046 (82%)] Loss: 0.135504 Make prediction for 5010 samples... 0.31562886 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 510 [0/25046 (0%)] Loss: 0.106127 Train epoch: 510 [329380/25046 (41%)] Loss: 0.144057 Train epoch: 510 [659160/25046 (82%)] Loss: 0.120582 Make prediction for 5010 samples... 0.28471592 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 511 [0/25046 (0%)] Loss: 0.122992 Train epoch: 511 [334840/25046 (41%)] Loss: 0.120110 Train epoch: 511 [652040/25046 (82%)] Loss: 0.163009 Make prediction for 5010 samples... 0.27841666 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 512 [0/25046 (0%)] Loss: 0.110318 Train epoch: 512 [333140/25046 (41%)] Loss: 0.136908 Train epoch: 512 [662960/25046 (82%)] Loss: 0.085537 Make prediction for 5010 samples... 0.2897582 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 513 [0/25046 (0%)] Loss: 0.103975 Train epoch: 513 [332200/25046 (41%)] Loss: 0.092721 Train epoch: 513 [660080/25046 (82%)] Loss: 0.138802 Make prediction for 5010 samples... 0.28212163 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 514 [0/25046 (0%)] Loss: 0.141488 Train epoch: 514 [331800/25046 (41%)] Loss: 0.122667 Train epoch: 514 [656760/25046 (82%)] Loss: 0.135793 Make prediction for 5010 samples... 0.2838559 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 515 [0/25046 (0%)] Loss: 0.146360 Train epoch: 515 [326560/25046 (41%)] Loss: 0.136040 Train epoch: 515 [658320/25046 (82%)] Loss: 0.185133 Make prediction for 5010 samples... 0.33129153 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 516 [0/25046 (0%)] Loss: 0.119501 Train epoch: 516 [331980/25046 (41%)] Loss: 0.124229 Train epoch: 516 [659040/25046 (82%)] Loss: 0.112891 Make prediction for 5010 samples... 0.2901171 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 517 [0/25046 (0%)] Loss: 0.114718 Train epoch: 517 [328100/25046 (41%)] Loss: 0.103807 Train epoch: 517 [660440/25046 (82%)] Loss: 0.121560 Make prediction for 5010 samples... 0.32949886 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 518 [0/25046 (0%)] Loss: 0.134746 Train epoch: 518 [331940/25046 (41%)] Loss: 0.138607 Train epoch: 518 [653160/25046 (82%)] Loss: 0.143116 Make prediction for 5010 samples... 0.2918617 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 519 [0/25046 (0%)] Loss: 0.109469 Train epoch: 519 [331040/25046 (41%)] Loss: 0.172808 Train epoch: 519 [654920/25046 (82%)] Loss: 0.172145 Make prediction for 5010 samples... 0.2776498 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 520 [0/25046 (0%)] Loss: 0.107921 Train epoch: 520 [330520/25046 (41%)] Loss: 0.139456 Train epoch: 520 [656280/25046 (82%)] Loss: 0.108979 Make prediction for 5010 samples... 0.29091588 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 521 [0/25046 (0%)] Loss: 0.087480 Train epoch: 521 [323440/25046 (41%)] Loss: 0.119898 Train epoch: 521 [648960/25046 (82%)] Loss: 0.118888 Make prediction for 5010 samples... 0.28607243 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 522 [0/25046 (0%)] Loss: 0.099046 Train epoch: 522 [328400/25046 (41%)] Loss: 0.105543 Train epoch: 522 [654840/25046 (82%)] Loss: 0.106340 Make prediction for 5010 samples... 0.28683594 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 523 [0/25046 (0%)] Loss: 0.091491 Train epoch: 523 [329600/25046 (41%)] Loss: 0.094045 Train epoch: 523 [663360/25046 (82%)] Loss: 0.114163 Make prediction for 5010 samples... 0.30172205 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 524 [0/25046 (0%)] Loss: 0.110748 Train epoch: 524 [326420/25046 (41%)] Loss: 0.146071 Train epoch: 524 [656080/25046 (82%)] Loss: 0.098246 Make prediction for 5010 samples... 0.29707924 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 525 [0/25046 (0%)] Loss: 0.116188 Train epoch: 525 [325500/25046 (41%)] Loss: 0.117001 Train epoch: 525 [648760/25046 (82%)] Loss: 0.139438 Make prediction for 5010 samples... 0.29234833 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 526 [0/25046 (0%)] Loss: 0.148087 Train epoch: 526 [327140/25046 (41%)] Loss: 0.146559 Train epoch: 526 [662640/25046 (82%)] Loss: 0.101842 Make prediction for 5010 samples... 0.2819699 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 527 [0/25046 (0%)] Loss: 0.126794 Train epoch: 527 [326720/25046 (41%)] Loss: 0.132368 Train epoch: 527 [663160/25046 (82%)] Loss: 0.208305 Make prediction for 5010 samples... 0.28733733 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 528 [0/25046 (0%)] Loss: 0.109429 Train epoch: 528 [330540/25046 (41%)] Loss: 0.122469 Train epoch: 528 [657840/25046 (82%)] Loss: 0.125843 Make prediction for 5010 samples... 0.28481668 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 529 [0/25046 (0%)] Loss: 0.146210 Train epoch: 529 [322040/25046 (41%)] Loss: 0.137711 Train epoch: 529 [664320/25046 (82%)] Loss: 0.179571 Make prediction for 5010 samples... 0.27976084 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 530 [0/25046 (0%)] Loss: 0.135042 Train epoch: 530 [326100/25046 (41%)] Loss: 0.112888 Train epoch: 530 [657320/25046 (82%)] Loss: 0.128461 Make prediction for 5010 samples... 0.28021967 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 531 [0/25046 (0%)] Loss: 0.117331 Train epoch: 531 [329320/25046 (41%)] Loss: 0.130758 Train epoch: 531 [665600/25046 (82%)] Loss: 0.106367 Make prediction for 5010 samples... 0.28074816 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 532 [0/25046 (0%)] Loss: 0.123252 Train epoch: 532 [325380/25046 (41%)] Loss: 0.083261 Train epoch: 532 [654160/25046 (82%)] Loss: 0.142760 Make prediction for 5010 samples... 0.28797683 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 533 [0/25046 (0%)] Loss: 0.119192 Train epoch: 533 [326680/25046 (41%)] Loss: 0.104324 Train epoch: 533 [658920/25046 (82%)] Loss: 0.146412 Make prediction for 5010 samples... 0.3226738 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 534 [0/25046 (0%)] Loss: 0.121372 Train epoch: 534 [329060/25046 (41%)] Loss: 0.143282 Train epoch: 534 [666480/25046 (82%)] Loss: 0.180729 Make prediction for 5010 samples... 0.30055568 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 535 [0/25046 (0%)] Loss: 0.156088 Train epoch: 535 [327680/25046 (41%)] Loss: 0.135903 Train epoch: 535 [650320/25046 (82%)] Loss: 0.155580 Make prediction for 5010 samples... 0.28355813 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 536 [0/25046 (0%)] Loss: 0.105849 Train epoch: 536 [324680/25046 (41%)] Loss: 0.153594 Train epoch: 536 [661560/25046 (82%)] Loss: 0.137564 Make prediction for 5010 samples... 0.27430367 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 537 [0/25046 (0%)] Loss: 0.089129 Train epoch: 537 [330140/25046 (41%)] Loss: 0.134688 Train epoch: 537 [660760/25046 (82%)] Loss: 0.144630 Make prediction for 5010 samples... 0.28253746 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 538 [0/25046 (0%)] Loss: 0.143113 Train epoch: 538 [324400/25046 (41%)] Loss: 0.141004 Train epoch: 538 [663680/25046 (82%)] Loss: 0.130647 Make prediction for 5010 samples... 0.28814492 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 539 [0/25046 (0%)] Loss: 0.129260 Train epoch: 539 [325600/25046 (41%)] Loss: 0.095603 Train epoch: 539 [655760/25046 (82%)] Loss: 0.127361 Make prediction for 5010 samples... 0.31342316 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 540 [0/25046 (0%)] Loss: 0.129789 Train epoch: 540 [328920/25046 (41%)] Loss: 0.098743 Train epoch: 540 [656000/25046 (82%)] Loss: 0.124297 Make prediction for 5010 samples... 0.27953988 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 541 [0/25046 (0%)] Loss: 0.104037 Train epoch: 541 [327800/25046 (41%)] Loss: 0.118263 Train epoch: 541 [655760/25046 (82%)] Loss: 0.119971 Make prediction for 5010 samples... 0.2845235 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 542 [0/25046 (0%)] Loss: 0.101374 Train epoch: 542 [329080/25046 (41%)] Loss: 0.124089 Train epoch: 542 [650440/25046 (82%)] Loss: 0.113565 Make prediction for 5010 samples... 0.28281274 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 543 [0/25046 (0%)] Loss: 0.107352 Train epoch: 543 [331840/25046 (41%)] Loss: 0.107475 Train epoch: 543 [656000/25046 (82%)] Loss: 0.128906 Make prediction for 5010 samples... 0.28690404 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 544 [0/25046 (0%)] Loss: 0.110287 Train epoch: 544 [331020/25046 (41%)] Loss: 0.108820 Train epoch: 544 [653160/25046 (82%)] Loss: 0.139855 Make prediction for 5010 samples... 0.29593274 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 545 [0/25046 (0%)] Loss: 0.120250 Train epoch: 545 [327500/25046 (41%)] Loss: 0.109799 Train epoch: 545 [661840/25046 (82%)] Loss: 0.124306 Make prediction for 5010 samples... 0.2809646 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 546 [0/25046 (0%)] Loss: 0.089917 Train epoch: 546 [328640/25046 (41%)] Loss: 0.106290 Train epoch: 546 [655480/25046 (82%)] Loss: 0.125640 Make prediction for 5010 samples... 0.3046674 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 547 [0/25046 (0%)] Loss: 0.127437 Train epoch: 547 [328540/25046 (41%)] Loss: 0.139676 Train epoch: 547 [653520/25046 (82%)] Loss: 0.165334 Make prediction for 5010 samples... 0.27874592 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 548 [0/25046 (0%)] Loss: 0.114963 Train epoch: 548 [327400/25046 (41%)] Loss: 0.135330 Train epoch: 548 [670240/25046 (82%)] Loss: 0.249341 Make prediction for 5010 samples... 0.29642144 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 549 [0/25046 (0%)] Loss: 0.160481 Train epoch: 549 [329680/25046 (41%)] Loss: 0.122459 Train epoch: 549 [657440/25046 (82%)] Loss: 0.130738 Make prediction for 5010 samples... 0.28865078 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 550 [0/25046 (0%)] Loss: 0.103013 Train epoch: 550 [325340/25046 (41%)] Loss: 0.128710 Train epoch: 550 [657680/25046 (82%)] Loss: 0.120599 Make prediction for 5010 samples... 0.308203 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 551 [0/25046 (0%)] Loss: 0.119604 Train epoch: 551 [330120/25046 (41%)] Loss: 0.106227 Train epoch: 551 [660040/25046 (82%)] Loss: 0.138159 Make prediction for 5010 samples... 0.3076018 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 552 [0/25046 (0%)] Loss: 0.094450 Train epoch: 552 [329060/25046 (41%)] Loss: 0.139633 Train epoch: 552 [660160/25046 (82%)] Loss: 0.204544 Make prediction for 5010 samples... 0.2837676 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 553 [0/25046 (0%)] Loss: 0.147755 Train epoch: 553 [333540/25046 (41%)] Loss: 0.114856 Train epoch: 553 [656480/25046 (82%)] Loss: 0.138059 Make prediction for 5010 samples... 0.29093584 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 554 [0/25046 (0%)] Loss: 0.120532 Train epoch: 554 [328280/25046 (41%)] Loss: 0.113436 Train epoch: 554 [648560/25046 (82%)] Loss: 0.133220 Make prediction for 5010 samples... 0.27657866 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 555 [0/25046 (0%)] Loss: 0.122444 Train epoch: 555 [327360/25046 (41%)] Loss: 0.101082 Train epoch: 555 [663840/25046 (82%)] Loss: 0.129007 Make prediction for 5010 samples... 0.2853306 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 556 [0/25046 (0%)] Loss: 0.103786 Train epoch: 556 [333360/25046 (41%)] Loss: 0.096469 Train epoch: 556 [666040/25046 (82%)] Loss: 0.119043 Make prediction for 5010 samples... 0.30054617 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 557 [0/25046 (0%)] Loss: 0.130073 Train epoch: 557 [327960/25046 (41%)] Loss: 0.093673 Train epoch: 557 [658080/25046 (82%)] Loss: 0.105098 Make prediction for 5010 samples... 0.29382464 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 558 [0/25046 (0%)] Loss: 0.104396 Train epoch: 558 [329900/25046 (41%)] Loss: 0.116078 Train epoch: 558 [663400/25046 (82%)] Loss: 0.110751 Make prediction for 5010 samples... 0.28060526 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 559 [0/25046 (0%)] Loss: 0.116169 Train epoch: 559 [330980/25046 (41%)] Loss: 0.147571 Train epoch: 559 [655040/25046 (82%)] Loss: 0.124844 Make prediction for 5010 samples... 0.30185297 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 560 [0/25046 (0%)] Loss: 0.130143 Train epoch: 560 [325700/25046 (41%)] Loss: 0.127459 Train epoch: 560 [651920/25046 (82%)] Loss: 0.105548 Make prediction for 5010 samples... 0.30850884 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 561 [0/25046 (0%)] Loss: 0.096186 Train epoch: 561 [331900/25046 (41%)] Loss: 0.104782 Train epoch: 561 [658440/25046 (82%)] Loss: 0.125708 Make prediction for 5010 samples... 0.2892056 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 562 [0/25046 (0%)] Loss: 0.129152 Train epoch: 562 [332040/25046 (41%)] Loss: 0.129146 Train epoch: 562 [657160/25046 (82%)] Loss: 0.114333 Make prediction for 5010 samples... 0.28504413 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 563 [0/25046 (0%)] Loss: 0.106049 Train epoch: 563 [329040/25046 (41%)] Loss: 0.108066 Train epoch: 563 [655640/25046 (82%)] Loss: 0.163642 Make prediction for 5010 samples... 0.2823066 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 564 [0/25046 (0%)] Loss: 0.117607 Train epoch: 564 [326860/25046 (41%)] Loss: 0.104843 Train epoch: 564 [653760/25046 (82%)] Loss: 0.131992 Make prediction for 5010 samples... 0.29113087 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 565 [0/25046 (0%)] Loss: 0.107103 Train epoch: 565 [325880/25046 (41%)] Loss: 0.107390 Train epoch: 565 [648160/25046 (82%)] Loss: 0.102423 Make prediction for 5010 samples... 0.27702132 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 566 [0/25046 (0%)] Loss: 0.105334 Train epoch: 566 [329080/25046 (41%)] Loss: 0.148128 Train epoch: 566 [660880/25046 (82%)] Loss: 0.115553 Make prediction for 5010 samples... 0.2880808 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 567 [0/25046 (0%)] Loss: 0.140213 Train epoch: 567 [327900/25046 (41%)] Loss: 0.137489 Train epoch: 567 [657120/25046 (82%)] Loss: 0.142714 Make prediction for 5010 samples... 0.28104892 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 568 [0/25046 (0%)] Loss: 0.103956 Train epoch: 568 [329220/25046 (41%)] Loss: 0.135440 Train epoch: 568 [654040/25046 (82%)] Loss: 0.104445 Make prediction for 5010 samples... 0.28705233 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 569 [0/25046 (0%)] Loss: 0.113406 Train epoch: 569 [321560/25046 (41%)] Loss: 0.113741 Train epoch: 569 [661360/25046 (82%)] Loss: 0.123582 Make prediction for 5010 samples... 0.27471235 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 570 [0/25046 (0%)] Loss: 0.118522 Train epoch: 570 [325240/25046 (41%)] Loss: 0.138069 Train epoch: 570 [647320/25046 (82%)] Loss: 0.109473 Make prediction for 5010 samples... 0.3164717 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 571 [0/25046 (0%)] Loss: 0.117713 Train epoch: 571 [330000/25046 (41%)] Loss: 0.136930 Train epoch: 571 [647880/25046 (82%)] Loss: 0.178875 Make prediction for 5010 samples... 0.28761533 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 572 [0/25046 (0%)] Loss: 0.130279 Train epoch: 572 [330760/25046 (41%)] Loss: 0.149070 Train epoch: 572 [655600/25046 (82%)] Loss: 0.101821 Make prediction for 5010 samples... 0.28469434 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 573 [0/25046 (0%)] Loss: 0.130488 Train epoch: 573 [324440/25046 (41%)] Loss: 0.105296 Train epoch: 573 [646480/25046 (82%)] Loss: 0.097591 Make prediction for 5010 samples... 0.27126133 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 574 [0/25046 (0%)] Loss: 0.096446 Train epoch: 574 [327700/25046 (41%)] Loss: 0.122587 Train epoch: 574 [649120/25046 (82%)] Loss: 0.148340 Make prediction for 5010 samples... 0.328109 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 575 [0/25046 (0%)] Loss: 0.110898 Train epoch: 575 [330080/25046 (41%)] Loss: 0.104429 Train epoch: 575 [662480/25046 (82%)] Loss: 0.127976 Make prediction for 5010 samples... 0.26919052 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 576 [0/25046 (0%)] Loss: 0.088543 Train epoch: 576 [331120/25046 (41%)] Loss: 0.110377 Train epoch: 576 [655600/25046 (82%)] Loss: 0.114777 Make prediction for 5010 samples... 0.29753 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 577 [0/25046 (0%)] Loss: 0.130209 Train epoch: 577 [325340/25046 (41%)] Loss: 0.140864 Train epoch: 577 [643480/25046 (82%)] Loss: 0.108414 Make prediction for 5010 samples... 0.30691016 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 578 [0/25046 (0%)] Loss: 0.113770 Train epoch: 578 [325000/25046 (41%)] Loss: 0.121079 Train epoch: 578 [661960/25046 (82%)] Loss: 0.103242 Make prediction for 5010 samples... 0.29132247 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 579 [0/25046 (0%)] Loss: 0.110920 Train epoch: 579 [322660/25046 (41%)] Loss: 0.110055 Train epoch: 579 [659160/25046 (82%)] Loss: 0.128479 Make prediction for 5010 samples... 0.2881722 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 580 [0/25046 (0%)] Loss: 0.105749 Train epoch: 580 [330080/25046 (41%)] Loss: 0.115210 Train epoch: 580 [651160/25046 (82%)] Loss: 0.103957 Make prediction for 5010 samples... 0.3434767 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 581 [0/25046 (0%)] Loss: 0.143297 Train epoch: 581 [328420/25046 (41%)] Loss: 0.138521 Train epoch: 581 [660520/25046 (82%)] Loss: 0.167192 Make prediction for 5010 samples... 0.29221332 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 582 [0/25046 (0%)] Loss: 0.113507 Train epoch: 582 [322420/25046 (41%)] Loss: 0.111832 Train epoch: 582 [657520/25046 (82%)] Loss: 0.124601 Make prediction for 5010 samples... 0.28933293 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 583 [0/25046 (0%)] Loss: 0.111381 Train epoch: 583 [329640/25046 (41%)] Loss: 0.136427 Train epoch: 583 [670480/25046 (82%)] Loss: 0.113388 Make prediction for 5010 samples... 0.28148353 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 584 [0/25046 (0%)] Loss: 0.085414 Train epoch: 584 [329620/25046 (41%)] Loss: 0.113373 Train epoch: 584 [660240/25046 (82%)] Loss: 0.115590 Make prediction for 5010 samples... 0.2883472 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 585 [0/25046 (0%)] Loss: 0.133239 Train epoch: 585 [323680/25046 (41%)] Loss: 0.089443 Train epoch: 585 [655680/25046 (82%)] Loss: 0.106641 Make prediction for 5010 samples... 0.2919734 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 586 [0/25046 (0%)] Loss: 0.117288 Train epoch: 586 [324160/25046 (41%)] Loss: 0.116473 Train epoch: 586 [654280/25046 (82%)] Loss: 0.112293 Make prediction for 5010 samples... 0.28924406 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 587 [0/25046 (0%)] Loss: 0.117529 Train epoch: 587 [327240/25046 (41%)] Loss: 0.117130 Train epoch: 587 [653000/25046 (82%)] Loss: 0.137184 Make prediction for 5010 samples... 0.30104488 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 588 [0/25046 (0%)] Loss: 0.112655 Train epoch: 588 [329240/25046 (41%)] Loss: 0.121500 Train epoch: 588 [658840/25046 (82%)] Loss: 0.144276 Make prediction for 5010 samples... 0.27690122 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 589 [0/25046 (0%)] Loss: 0.132194 Train epoch: 589 [326120/25046 (41%)] Loss: 0.108692 Train epoch: 589 [664720/25046 (82%)] Loss: 0.106747 Make prediction for 5010 samples... 0.31147495 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 590 [0/25046 (0%)] Loss: 0.108306 Train epoch: 590 [323240/25046 (41%)] Loss: 0.142300 Train epoch: 590 [655840/25046 (82%)] Loss: 0.117178 Make prediction for 5010 samples... 0.2826842 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 591 [0/25046 (0%)] Loss: 0.115251 Train epoch: 591 [328280/25046 (41%)] Loss: 0.096561 Train epoch: 591 [657560/25046 (82%)] Loss: 0.135510 Make prediction for 5010 samples... 0.28548294 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 592 [0/25046 (0%)] Loss: 0.134368 Train epoch: 592 [326440/25046 (41%)] Loss: 0.124737 Train epoch: 592 [660480/25046 (82%)] Loss: 0.100989 Make prediction for 5010 samples... 0.27512452 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 593 [0/25046 (0%)] Loss: 0.096778 Train epoch: 593 [330720/25046 (41%)] Loss: 0.116846 Train epoch: 593 [656120/25046 (82%)] Loss: 0.103957 Make prediction for 5010 samples... 0.30354348 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 594 [0/25046 (0%)] Loss: 0.114752 Train epoch: 594 [325900/25046 (41%)] Loss: 0.116095 Train epoch: 594 [654560/25046 (82%)] Loss: 0.117451 Make prediction for 5010 samples... 0.279987 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 595 [0/25046 (0%)] Loss: 0.081155 Train epoch: 595 [325120/25046 (41%)] Loss: 0.132536 Train epoch: 595 [657960/25046 (82%)] Loss: 0.127564 Make prediction for 5010 samples... 0.29728264 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 596 [0/25046 (0%)] Loss: 0.110860 Train epoch: 596 [326020/25046 (41%)] Loss: 0.119086 Train epoch: 596 [653200/25046 (82%)] Loss: 0.120403 Make prediction for 5010 samples... 0.28117627 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 597 [0/25046 (0%)] Loss: 0.114581 Train epoch: 597 [330240/25046 (41%)] Loss: 0.143189 Train epoch: 597 [662960/25046 (82%)] Loss: 0.140337 Make prediction for 5010 samples... 0.281507 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 598 [0/25046 (0%)] Loss: 0.085791 Train epoch: 598 [323300/25046 (41%)] Loss: 0.119625 Train epoch: 598 [656360/25046 (82%)] Loss: 0.127705 Make prediction for 5010 samples... 0.28038707 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 599 [0/25046 (0%)] Loss: 0.094227 Train epoch: 599 [331500/25046 (41%)] Loss: 0.130517 Train epoch: 599 [650880/25046 (82%)] Loss: 0.110864 Make prediction for 5010 samples... 0.27509195 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 600 [0/25046 (0%)] Loss: 0.089515 Train epoch: 600 [328540/25046 (41%)] Loss: 0.153087 Train epoch: 600 [665080/25046 (82%)] Loss: 0.200077 Make prediction for 5010 samples... 0.29663882 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 601 [0/25046 (0%)] Loss: 0.111344 Train epoch: 601 [324100/25046 (41%)] Loss: 0.158197 Train epoch: 601 [662200/25046 (82%)] Loss: 0.135891 Make prediction for 5010 samples... 0.27846828 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 602 [0/25046 (0%)] Loss: 0.127437 Train epoch: 602 [321700/25046 (41%)] Loss: 0.100208 Train epoch: 602 [653960/25046 (82%)] Loss: 0.091085 Make prediction for 5010 samples... 0.28368935 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 603 [0/25046 (0%)] Loss: 0.124165 Train epoch: 603 [326280/25046 (41%)] Loss: 0.156493 Train epoch: 603 [657120/25046 (82%)] Loss: 0.127283 Make prediction for 5010 samples... 0.2973682 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 604 [0/25046 (0%)] Loss: 0.129156 Train epoch: 604 [332980/25046 (41%)] Loss: 0.095019 Train epoch: 604 [651080/25046 (82%)] Loss: 0.094815 Make prediction for 5010 samples... 0.27883145 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 605 [0/25046 (0%)] Loss: 0.129740 Train epoch: 605 [329000/25046 (41%)] Loss: 0.088479 Train epoch: 605 [662080/25046 (82%)] Loss: 0.117772 Make prediction for 5010 samples... 0.27210334 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 606 [0/25046 (0%)] Loss: 0.114167 Train epoch: 606 [327060/25046 (41%)] Loss: 0.099817 Train epoch: 606 [657600/25046 (82%)] Loss: 0.118510 Make prediction for 5010 samples... 0.2758221 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 607 [0/25046 (0%)] Loss: 0.162518 Train epoch: 607 [333640/25046 (41%)] Loss: 0.090204 Train epoch: 607 [652400/25046 (82%)] Loss: 0.147643 Make prediction for 5010 samples... 0.40208778 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 608 [0/25046 (0%)] Loss: 0.180866 Train epoch: 608 [329280/25046 (41%)] Loss: 0.157152 Train epoch: 608 [666440/25046 (82%)] Loss: 0.116461 Make prediction for 5010 samples... 0.2906192 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 609 [0/25046 (0%)] Loss: 0.121270 Train epoch: 609 [324140/25046 (41%)] Loss: 0.139715 Train epoch: 609 [653360/25046 (82%)] Loss: 0.123175 Make prediction for 5010 samples... 0.2754894 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 610 [0/25046 (0%)] Loss: 0.115597 Train epoch: 610 [328640/25046 (41%)] Loss: 0.092327 Train epoch: 610 [652000/25046 (82%)] Loss: 0.128427 Make prediction for 5010 samples... 0.28531447 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 611 [0/25046 (0%)] Loss: 0.095976 Train epoch: 611 [326940/25046 (41%)] Loss: 0.107616 Train epoch: 611 [649360/25046 (82%)] Loss: 0.122466 Make prediction for 5010 samples... 0.28556928 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 612 [0/25046 (0%)] Loss: 0.136469 Train epoch: 612 [332580/25046 (41%)] Loss: 0.107599 Train epoch: 612 [649840/25046 (82%)] Loss: 0.119619 Make prediction for 5010 samples... 0.28067124 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 613 [0/25046 (0%)] Loss: 0.101907 Train epoch: 613 [328460/25046 (41%)] Loss: 0.117275 Train epoch: 613 [656440/25046 (82%)] Loss: 0.118572 Make prediction for 5010 samples... 0.27951708 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 614 [0/25046 (0%)] Loss: 0.106801 Train epoch: 614 [321520/25046 (41%)] Loss: 0.101326 Train epoch: 614 [650480/25046 (82%)] Loss: 0.094277 Make prediction for 5010 samples... 0.280286 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 615 [0/25046 (0%)] Loss: 0.112170 Train epoch: 615 [327540/25046 (41%)] Loss: 0.117867 Train epoch: 615 [654200/25046 (82%)] Loss: 0.114971 Make prediction for 5010 samples... 0.2846496 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 616 [0/25046 (0%)] Loss: 0.094188 Train epoch: 616 [325200/25046 (41%)] Loss: 0.155688 Train epoch: 616 [657520/25046 (82%)] Loss: 0.138321 Make prediction for 5010 samples... 0.27449095 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 617 [0/25046 (0%)] Loss: 0.103658 Train epoch: 617 [329340/25046 (41%)] Loss: 0.143712 Train epoch: 617 [660720/25046 (82%)] Loss: 0.101742 Make prediction for 5010 samples... 0.27226663 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 618 [0/25046 (0%)] Loss: 0.099397 Train epoch: 618 [330300/25046 (41%)] Loss: 0.101462 Train epoch: 618 [661080/25046 (82%)] Loss: 0.122924 Make prediction for 5010 samples... 0.3061365 No improvement since epoch 422 ; best_mse,best_ci: 0.26857442 0.8765902031615915 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 619 [0/25046 (0%)] Loss: 0.136486 Train epoch: 619 [328820/25046 (41%)] Loss: 0.109710 Train epoch: 619 [651360/25046 (82%)] Loss: 0.105957 Make prediction for 5010 samples... rmse improved at epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 620 [0/25046 (0%)] Loss: 0.098999 Train epoch: 620 [329800/25046 (41%)] Loss: 0.118866 Train epoch: 620 [655800/25046 (82%)] Loss: 0.090964 Make prediction for 5010 samples... 0.2888855 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 621 [0/25046 (0%)] Loss: 0.113357 Train epoch: 621 [335140/25046 (41%)] Loss: 0.105000 Train epoch: 621 [652400/25046 (82%)] Loss: 0.111800 Make prediction for 5010 samples... 0.28188705 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 622 [0/25046 (0%)] Loss: 0.144613 Train epoch: 622 [327860/25046 (41%)] Loss: 0.109147 Train epoch: 622 [664040/25046 (82%)] Loss: 0.174837 Make prediction for 5010 samples... 0.32357666 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 623 [0/25046 (0%)] Loss: 0.139412 Train epoch: 623 [327900/25046 (41%)] Loss: 0.115584 Train epoch: 623 [648880/25046 (82%)] Loss: 0.116295 Make prediction for 5010 samples... 0.2925436 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 624 [0/25046 (0%)] Loss: 0.152298 Train epoch: 624 [329860/25046 (41%)] Loss: 0.103263 Train epoch: 624 [660760/25046 (82%)] Loss: 0.117118 Make prediction for 5010 samples... 0.27497402 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 625 [0/25046 (0%)] Loss: 0.116822 Train epoch: 625 [328800/25046 (41%)] Loss: 0.141609 Train epoch: 625 [653320/25046 (82%)] Loss: 0.134564 Make prediction for 5010 samples... 0.28224915 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 626 [0/25046 (0%)] Loss: 0.107067 Train epoch: 626 [331900/25046 (41%)] Loss: 0.085286 Train epoch: 626 [654360/25046 (82%)] Loss: 0.107237 Make prediction for 5010 samples... 0.27547795 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 627 [0/25046 (0%)] Loss: 0.103915 Train epoch: 627 [328900/25046 (41%)] Loss: 0.098114 Train epoch: 627 [661240/25046 (82%)] Loss: 0.103298 Make prediction for 5010 samples... 0.3477008 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 628 [0/25046 (0%)] Loss: 0.134294 Train epoch: 628 [329080/25046 (41%)] Loss: 0.112109 Train epoch: 628 [655720/25046 (82%)] Loss: 0.131484 Make prediction for 5010 samples... 0.27370226 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 629 [0/25046 (0%)] Loss: 0.130578 Train epoch: 629 [329860/25046 (41%)] Loss: 0.110048 Train epoch: 629 [665760/25046 (82%)] Loss: 0.100239 Make prediction for 5010 samples... 0.27950808 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 630 [0/25046 (0%)] Loss: 0.096856 Train epoch: 630 [332860/25046 (41%)] Loss: 0.119463 Train epoch: 630 [668440/25046 (82%)] Loss: 0.138528 Make prediction for 5010 samples... 0.27518952 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 631 [0/25046 (0%)] Loss: 0.117332 Train epoch: 631 [325960/25046 (41%)] Loss: 0.082603 Train epoch: 631 [652520/25046 (82%)] Loss: 0.111846 Make prediction for 5010 samples... 0.347033 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 632 [0/25046 (0%)] Loss: 0.153272 Train epoch: 632 [325120/25046 (41%)] Loss: 0.128022 Train epoch: 632 [653840/25046 (82%)] Loss: 0.095379 Make prediction for 5010 samples... 0.32766303 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 633 [0/25046 (0%)] Loss: 0.124815 Train epoch: 633 [331320/25046 (41%)] Loss: 0.120324 Train epoch: 633 [660680/25046 (82%)] Loss: 0.103259 Make prediction for 5010 samples... 0.36502114 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 634 [0/25046 (0%)] Loss: 0.133164 Train epoch: 634 [329180/25046 (41%)] Loss: 0.124292 Train epoch: 634 [658120/25046 (82%)] Loss: 0.116931 Make prediction for 5010 samples... 0.27463603 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 635 [0/25046 (0%)] Loss: 0.090345 Train epoch: 635 [327760/25046 (41%)] Loss: 0.100059 Train epoch: 635 [659960/25046 (82%)] Loss: 0.127276 Make prediction for 5010 samples... 0.27455118 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 636 [0/25046 (0%)] Loss: 0.117839 Train epoch: 636 [327260/25046 (41%)] Loss: 0.130436 Train epoch: 636 [650760/25046 (82%)] Loss: 0.122912 Make prediction for 5010 samples... 0.31025484 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 637 [0/25046 (0%)] Loss: 0.131848 Train epoch: 637 [330240/25046 (41%)] Loss: 0.111516 Train epoch: 637 [663000/25046 (82%)] Loss: 0.117016 Make prediction for 5010 samples... 0.31862977 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 638 [0/25046 (0%)] Loss: 0.122319 Train epoch: 638 [325340/25046 (41%)] Loss: 0.117629 Train epoch: 638 [659680/25046 (82%)] Loss: 0.113422 Make prediction for 5010 samples... 0.26647583 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 639 [0/25046 (0%)] Loss: 0.115193 Train epoch: 639 [328660/25046 (41%)] Loss: 0.097172 Train epoch: 639 [649560/25046 (82%)] Loss: 0.100133 Make prediction for 5010 samples... 0.2896006 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 640 [0/25046 (0%)] Loss: 0.103823 Train epoch: 640 [333400/25046 (41%)] Loss: 0.117848 Train epoch: 640 [667040/25046 (82%)] Loss: 0.102278 Make prediction for 5010 samples... 0.30662894 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 641 [0/25046 (0%)] Loss: 0.120085 Train epoch: 641 [327360/25046 (41%)] Loss: 0.125645 Train epoch: 641 [665160/25046 (82%)] Loss: 0.102157 Make prediction for 5010 samples... 0.28523162 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 642 [0/25046 (0%)] Loss: 0.111587 Train epoch: 642 [332540/25046 (41%)] Loss: 0.125643 Train epoch: 642 [649480/25046 (82%)] Loss: 0.116875 Make prediction for 5010 samples... 0.2943255 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 643 [0/25046 (0%)] Loss: 0.097282 Train epoch: 643 [326420/25046 (41%)] Loss: 0.144114 Train epoch: 643 [658760/25046 (82%)] Loss: 0.091940 Make prediction for 5010 samples... 0.28100175 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 644 [0/25046 (0%)] Loss: 0.120616 Train epoch: 644 [330980/25046 (41%)] Loss: 0.103243 Train epoch: 644 [662600/25046 (82%)] Loss: 0.125057 Make prediction for 5010 samples... 0.27453154 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 645 [0/25046 (0%)] Loss: 0.098599 Train epoch: 645 [333160/25046 (41%)] Loss: 0.140020 Train epoch: 645 [650720/25046 (82%)] Loss: 0.118134 Make prediction for 5010 samples... 0.30099297 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 646 [0/25046 (0%)] Loss: 0.132516 Train epoch: 646 [327660/25046 (41%)] Loss: 0.102629 Train epoch: 646 [663200/25046 (82%)] Loss: 0.121468 Make prediction for 5010 samples... 0.27012542 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 647 [0/25046 (0%)] Loss: 0.121365 Train epoch: 647 [327420/25046 (41%)] Loss: 0.144570 Train epoch: 647 [660240/25046 (82%)] Loss: 0.111761 Make prediction for 5010 samples... 0.286869 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 648 [0/25046 (0%)] Loss: 0.119453 Train epoch: 648 [329540/25046 (41%)] Loss: 0.135602 Train epoch: 648 [661400/25046 (82%)] Loss: 0.104543 Make prediction for 5010 samples... 0.27902663 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 649 [0/25046 (0%)] Loss: 0.091038 Train epoch: 649 [326400/25046 (41%)] Loss: 0.104349 Train epoch: 649 [645880/25046 (82%)] Loss: 0.150412 Make prediction for 5010 samples... 0.30243543 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 650 [0/25046 (0%)] Loss: 0.086518 Train epoch: 650 [329800/25046 (41%)] Loss: 0.091147 Train epoch: 650 [658880/25046 (82%)] Loss: 0.082537 Make prediction for 5010 samples... 0.27137774 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 651 [0/25046 (0%)] Loss: 0.114206 Train epoch: 651 [326040/25046 (41%)] Loss: 0.123668 Train epoch: 651 [652680/25046 (82%)] Loss: 0.096101 Make prediction for 5010 samples... 0.2708892 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 652 [0/25046 (0%)] Loss: 0.110794 Train epoch: 652 [329820/25046 (41%)] Loss: 0.110377 Train epoch: 652 [657320/25046 (82%)] Loss: 0.117234 Make prediction for 5010 samples... 0.28706014 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 653 [0/25046 (0%)] Loss: 0.118386 Train epoch: 653 [334520/25046 (41%)] Loss: 0.116923 Train epoch: 653 [659480/25046 (82%)] Loss: 0.106623 Make prediction for 5010 samples... 0.3104525 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 654 [0/25046 (0%)] Loss: 0.112997 Train epoch: 654 [327280/25046 (41%)] Loss: 0.099371 Train epoch: 654 [671240/25046 (82%)] Loss: 0.132739 Make prediction for 5010 samples... 0.28046185 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 655 [0/25046 (0%)] Loss: 0.094613 Train epoch: 655 [323760/25046 (41%)] Loss: 0.099352 Train epoch: 655 [655520/25046 (82%)] Loss: 0.128852 Make prediction for 5010 samples... 0.33844042 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 656 [0/25046 (0%)] Loss: 0.122065 Train epoch: 656 [332520/25046 (41%)] Loss: 0.125560 Train epoch: 656 [655680/25046 (82%)] Loss: 0.099777 Make prediction for 5010 samples... 0.3103635 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 657 [0/25046 (0%)] Loss: 0.135469 Train epoch: 657 [325860/25046 (41%)] Loss: 0.133366 Train epoch: 657 [648560/25046 (82%)] Loss: 0.090217 Make prediction for 5010 samples... 0.2703724 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 658 [0/25046 (0%)] Loss: 0.088871 Train epoch: 658 [327340/25046 (41%)] Loss: 0.123641 Train epoch: 658 [655480/25046 (82%)] Loss: 0.126315 Make prediction for 5010 samples... 0.290381 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 659 [0/25046 (0%)] Loss: 0.100862 Train epoch: 659 [326460/25046 (41%)] Loss: 0.131225 Train epoch: 659 [654680/25046 (82%)] Loss: 0.095398 Make prediction for 5010 samples... 0.28276348 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 660 [0/25046 (0%)] Loss: 0.081977 Train epoch: 660 [328380/25046 (41%)] Loss: 0.101941 Train epoch: 660 [643560/25046 (82%)] Loss: 0.084708 Make prediction for 5010 samples... 0.28016287 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 661 [0/25046 (0%)] Loss: 0.103666 Train epoch: 661 [330440/25046 (41%)] Loss: 0.098968 Train epoch: 661 [657480/25046 (82%)] Loss: 0.165387 Make prediction for 5010 samples... 0.29229018 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 662 [0/25046 (0%)] Loss: 0.132178 Train epoch: 662 [326080/25046 (41%)] Loss: 0.136098 Train epoch: 662 [648040/25046 (82%)] Loss: 0.089208 Make prediction for 5010 samples... 0.27982116 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 663 [0/25046 (0%)] Loss: 0.112912 Train epoch: 663 [327460/25046 (41%)] Loss: 0.108389 Train epoch: 663 [658240/25046 (82%)] Loss: 0.134973 Make prediction for 5010 samples... 0.27443442 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 664 [0/25046 (0%)] Loss: 0.104650 Train epoch: 664 [325100/25046 (41%)] Loss: 0.098501 Train epoch: 664 [657640/25046 (82%)] Loss: 0.106303 Make prediction for 5010 samples... 0.2989078 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 665 [0/25046 (0%)] Loss: 0.131277 Train epoch: 665 [327660/25046 (41%)] Loss: 0.131753 Train epoch: 665 [657080/25046 (82%)] Loss: 0.116960 Make prediction for 5010 samples... 0.2895626 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 666 [0/25046 (0%)] Loss: 0.104415 Train epoch: 666 [327240/25046 (41%)] Loss: 0.106226 Train epoch: 666 [655200/25046 (82%)] Loss: 0.092749 Make prediction for 5010 samples... 0.28757328 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 667 [0/25046 (0%)] Loss: 0.108990 Train epoch: 667 [328360/25046 (41%)] Loss: 0.134151 Train epoch: 667 [664480/25046 (82%)] Loss: 0.088978 Make prediction for 5010 samples... 0.27929786 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 668 [0/25046 (0%)] Loss: 0.071197 Train epoch: 668 [334040/25046 (41%)] Loss: 0.119701 Train epoch: 668 [653720/25046 (82%)] Loss: 0.126864 Make prediction for 5010 samples... 0.31733984 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 669 [0/25046 (0%)] Loss: 0.132600 Train epoch: 669 [324480/25046 (41%)] Loss: 0.116827 Train epoch: 669 [663800/25046 (82%)] Loss: 0.101304 Make prediction for 5010 samples... 0.29596102 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 670 [0/25046 (0%)] Loss: 0.103485 Train epoch: 670 [328220/25046 (41%)] Loss: 0.122727 Train epoch: 670 [664120/25046 (82%)] Loss: 0.116607 Make prediction for 5010 samples... 0.3272169 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 671 [0/25046 (0%)] Loss: 0.124313 Train epoch: 671 [327600/25046 (41%)] Loss: 0.152860 Train epoch: 671 [653000/25046 (82%)] Loss: 0.134926 Make prediction for 5010 samples... 0.28377062 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 672 [0/25046 (0%)] Loss: 0.108955 Train epoch: 672 [326040/25046 (41%)] Loss: 0.131685 Train epoch: 672 [653400/25046 (82%)] Loss: 0.092094 Make prediction for 5010 samples... 0.28955522 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 673 [0/25046 (0%)] Loss: 0.105634 Train epoch: 673 [326940/25046 (41%)] Loss: 0.114075 Train epoch: 673 [665720/25046 (82%)] Loss: 0.134763 Make prediction for 5010 samples... 0.2891632 No improvement since epoch 619 ; best_mse,best_ci: 0.26613593 0.8809833870720475 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 674 [0/25046 (0%)] Loss: 0.115244 Train epoch: 674 [324500/25046 (41%)] Loss: 0.084036 Train epoch: 674 [650320/25046 (82%)] Loss: 0.107122 Make prediction for 5010 samples... rmse improved at epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 675 [0/25046 (0%)] Loss: 0.100856 Train epoch: 675 [327940/25046 (41%)] Loss: 0.102521 Train epoch: 675 [664920/25046 (82%)] Loss: 0.093187 Make prediction for 5010 samples... 0.28234604 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 676 [0/25046 (0%)] Loss: 0.104027 Train epoch: 676 [328040/25046 (41%)] Loss: 0.134971 Train epoch: 676 [652720/25046 (82%)] Loss: 0.133293 Make prediction for 5010 samples... 0.2986694 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 677 [0/25046 (0%)] Loss: 0.131297 Train epoch: 677 [330140/25046 (41%)] Loss: 0.120933 Train epoch: 677 [650240/25046 (82%)] Loss: 0.122919 Make prediction for 5010 samples... 0.29785722 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 678 [0/25046 (0%)] Loss: 0.101925 Train epoch: 678 [328540/25046 (41%)] Loss: 0.152232 Train epoch: 678 [658560/25046 (82%)] Loss: 0.138389 Make prediction for 5010 samples... 0.29702118 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 679 [0/25046 (0%)] Loss: 0.101861 Train epoch: 679 [326440/25046 (41%)] Loss: 0.114136 Train epoch: 679 [651880/25046 (82%)] Loss: 0.121158 Make prediction for 5010 samples... 0.27336022 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 680 [0/25046 (0%)] Loss: 0.132145 Train epoch: 680 [328040/25046 (41%)] Loss: 0.115750 Train epoch: 680 [661040/25046 (82%)] Loss: 0.087501 Make prediction for 5010 samples... 0.29217324 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 681 [0/25046 (0%)] Loss: 0.108310 Train epoch: 681 [328680/25046 (41%)] Loss: 0.090679 Train epoch: 681 [654760/25046 (82%)] Loss: 0.143649 Make prediction for 5010 samples... 0.27957815 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 682 [0/25046 (0%)] Loss: 0.113692 Train epoch: 682 [325300/25046 (41%)] Loss: 0.099336 Train epoch: 682 [662480/25046 (82%)] Loss: 0.128379 Make prediction for 5010 samples... 0.27734205 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 683 [0/25046 (0%)] Loss: 0.136003 Train epoch: 683 [330240/25046 (41%)] Loss: 0.098304 Train epoch: 683 [662960/25046 (82%)] Loss: 0.119677 Make prediction for 5010 samples... 0.28463772 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 684 [0/25046 (0%)] Loss: 0.115006 Train epoch: 684 [328360/25046 (41%)] Loss: 0.102095 Train epoch: 684 [653360/25046 (82%)] Loss: 0.112212 Make prediction for 5010 samples... 0.3000155 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 685 [0/25046 (0%)] Loss: 0.139135 Train epoch: 685 [332960/25046 (41%)] Loss: 0.102831 Train epoch: 685 [657400/25046 (82%)] Loss: 0.104466 Make prediction for 5010 samples... 0.28275573 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 686 [0/25046 (0%)] Loss: 0.102457 Train epoch: 686 [333240/25046 (41%)] Loss: 0.107712 Train epoch: 686 [653320/25046 (82%)] Loss: 0.091266 Make prediction for 5010 samples... 0.29315227 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 687 [0/25046 (0%)] Loss: 0.088509 Train epoch: 687 [327180/25046 (41%)] Loss: 0.108706 Train epoch: 687 [662640/25046 (82%)] Loss: 0.094973 Make prediction for 5010 samples... 0.29434577 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 688 [0/25046 (0%)] Loss: 0.085965 Train epoch: 688 [330320/25046 (41%)] Loss: 0.100006 Train epoch: 688 [659520/25046 (82%)] Loss: 0.091229 Make prediction for 5010 samples... 0.27830863 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 689 [0/25046 (0%)] Loss: 0.132695 Train epoch: 689 [326140/25046 (41%)] Loss: 0.126499 Train epoch: 689 [655320/25046 (82%)] Loss: 0.097088 Make prediction for 5010 samples... 0.27821058 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 690 [0/25046 (0%)] Loss: 0.086665 Train epoch: 690 [328900/25046 (41%)] Loss: 0.155157 Train epoch: 690 [656400/25046 (82%)] Loss: 0.087410 Make prediction for 5010 samples... 0.31782445 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 691 [0/25046 (0%)] Loss: 0.124503 Train epoch: 691 [325800/25046 (41%)] Loss: 0.083838 Train epoch: 691 [652400/25046 (82%)] Loss: 0.093310 Make prediction for 5010 samples... 0.2726992 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 692 [0/25046 (0%)] Loss: 0.090230 Train epoch: 692 [327620/25046 (41%)] Loss: 0.139911 Train epoch: 692 [658720/25046 (82%)] Loss: 0.115468 Make prediction for 5010 samples... 0.2832999 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 693 [0/25046 (0%)] Loss: 0.099589 Train epoch: 693 [328060/25046 (41%)] Loss: 0.122102 Train epoch: 693 [656240/25046 (82%)] Loss: 0.103849 Make prediction for 5010 samples... 0.29272884 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 694 [0/25046 (0%)] Loss: 0.119877 Train epoch: 694 [332520/25046 (41%)] Loss: 0.112249 Train epoch: 694 [661840/25046 (82%)] Loss: 0.147899 Make prediction for 5010 samples... 0.28969595 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 695 [0/25046 (0%)] Loss: 0.114156 Train epoch: 695 [327520/25046 (41%)] Loss: 0.136952 Train epoch: 695 [656680/25046 (82%)] Loss: 0.087736 Make prediction for 5010 samples... 0.27490205 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 696 [0/25046 (0%)] Loss: 0.102642 Train epoch: 696 [326920/25046 (41%)] Loss: 0.101506 Train epoch: 696 [651640/25046 (82%)] Loss: 0.093152 Make prediction for 5010 samples... 0.2831937 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 697 [0/25046 (0%)] Loss: 0.088210 Train epoch: 697 [325000/25046 (41%)] Loss: 0.079986 Train epoch: 697 [657320/25046 (82%)] Loss: 0.102181 Make prediction for 5010 samples... 0.30129912 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 698 [0/25046 (0%)] Loss: 0.179978 Train epoch: 698 [324180/25046 (41%)] Loss: 0.112557 Train epoch: 698 [657040/25046 (82%)] Loss: 0.094110 Make prediction for 5010 samples... 0.32678464 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 699 [0/25046 (0%)] Loss: 0.090224 Train epoch: 699 [332220/25046 (41%)] Loss: 0.113694 Train epoch: 699 [655480/25046 (82%)] Loss: 0.142294 Make prediction for 5010 samples... 0.29898 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 700 [0/25046 (0%)] Loss: 0.101256 Train epoch: 700 [325580/25046 (41%)] Loss: 0.104455 Train epoch: 700 [659640/25046 (82%)] Loss: 0.123793 Make prediction for 5010 samples... 0.28378043 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 701 [0/25046 (0%)] Loss: 0.094657 Train epoch: 701 [329920/25046 (41%)] Loss: 0.077062 Train epoch: 701 [657600/25046 (82%)] Loss: 0.086342 Make prediction for 5010 samples... 0.2862875 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 702 [0/25046 (0%)] Loss: 0.106889 Train epoch: 702 [326160/25046 (41%)] Loss: 0.186450 Train epoch: 702 [660000/25046 (82%)] Loss: 0.112840 Make prediction for 5010 samples... 0.28217208 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 703 [0/25046 (0%)] Loss: 0.096569 Train epoch: 703 [332200/25046 (41%)] Loss: 0.110047 Train epoch: 703 [661640/25046 (82%)] Loss: 0.113731 Make prediction for 5010 samples... 0.28163433 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 704 [0/25046 (0%)] Loss: 0.123161 Train epoch: 704 [327560/25046 (41%)] Loss: 0.104805 Train epoch: 704 [659400/25046 (82%)] Loss: 0.116838 Make prediction for 5010 samples... 0.27507496 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 705 [0/25046 (0%)] Loss: 0.100151 Train epoch: 705 [330140/25046 (41%)] Loss: 0.094454 Train epoch: 705 [657280/25046 (82%)] Loss: 0.114034 Make prediction for 5010 samples... 0.29871103 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 706 [0/25046 (0%)] Loss: 0.101837 Train epoch: 706 [324980/25046 (41%)] Loss: 0.106276 Train epoch: 706 [654200/25046 (82%)] Loss: 0.104691 Make prediction for 5010 samples... 0.27839553 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 707 [0/25046 (0%)] Loss: 0.069768 Train epoch: 707 [326820/25046 (41%)] Loss: 0.123918 Train epoch: 707 [659480/25046 (82%)] Loss: 0.102983 Make prediction for 5010 samples... 0.283578 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 708 [0/25046 (0%)] Loss: 0.108468 Train epoch: 708 [322500/25046 (41%)] Loss: 0.078026 Train epoch: 708 [653840/25046 (82%)] Loss: 0.148498 Make prediction for 5010 samples... 0.2843671 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 709 [0/25046 (0%)] Loss: 0.096620 Train epoch: 709 [334200/25046 (41%)] Loss: 0.100166 Train epoch: 709 [650280/25046 (82%)] Loss: 0.092125 Make prediction for 5010 samples... 0.2775103 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 710 [0/25046 (0%)] Loss: 0.083003 Train epoch: 710 [326660/25046 (41%)] Loss: 0.104649 Train epoch: 710 [659280/25046 (82%)] Loss: 0.092547 Make prediction for 5010 samples... 0.2877461 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 711 [0/25046 (0%)] Loss: 0.082844 Train epoch: 711 [323500/25046 (41%)] Loss: 0.178475 Train epoch: 711 [657200/25046 (82%)] Loss: 0.076915 Make prediction for 5010 samples... 0.28079838 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 712 [0/25046 (0%)] Loss: 0.101185 Train epoch: 712 [329080/25046 (41%)] Loss: 0.106288 Train epoch: 712 [641560/25046 (82%)] Loss: 0.089364 Make prediction for 5010 samples... 0.27008542 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 713 [0/25046 (0%)] Loss: 0.107404 Train epoch: 713 [326220/25046 (41%)] Loss: 0.134371 Train epoch: 713 [661960/25046 (82%)] Loss: 0.105338 Make prediction for 5010 samples... 0.29504535 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 714 [0/25046 (0%)] Loss: 0.069065 Train epoch: 714 [332080/25046 (41%)] Loss: 0.141039 Train epoch: 714 [663520/25046 (82%)] Loss: 0.106081 Make prediction for 5010 samples... 0.27703586 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 715 [0/25046 (0%)] Loss: 0.107675 Train epoch: 715 [326180/25046 (41%)] Loss: 0.156614 Train epoch: 715 [650160/25046 (82%)] Loss: 0.107343 Make prediction for 5010 samples... 0.31641287 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 716 [0/25046 (0%)] Loss: 0.111952 Train epoch: 716 [331260/25046 (41%)] Loss: 0.105768 Train epoch: 716 [657040/25046 (82%)] Loss: 0.122896 Make prediction for 5010 samples... 0.26533416 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 717 [0/25046 (0%)] Loss: 0.101404 Train epoch: 717 [325560/25046 (41%)] Loss: 0.112465 Train epoch: 717 [666120/25046 (82%)] Loss: 0.105640 Make prediction for 5010 samples... 0.2956044 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 718 [0/25046 (0%)] Loss: 0.076889 Train epoch: 718 [326200/25046 (41%)] Loss: 0.092249 Train epoch: 718 [655160/25046 (82%)] Loss: 0.103966 Make prediction for 5010 samples... 0.28024825 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 719 [0/25046 (0%)] Loss: 0.137026 Train epoch: 719 [325680/25046 (41%)] Loss: 0.097236 Train epoch: 719 [657000/25046 (82%)] Loss: 0.117930 Make prediction for 5010 samples... 0.28938743 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 720 [0/25046 (0%)] Loss: 0.113207 Train epoch: 720 [327680/25046 (41%)] Loss: 0.076743 Train epoch: 720 [659600/25046 (82%)] Loss: 0.098354 Make prediction for 5010 samples... 0.26974034 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 721 [0/25046 (0%)] Loss: 0.108831 Train epoch: 721 [326660/25046 (41%)] Loss: 0.093252 Train epoch: 721 [650440/25046 (82%)] Loss: 0.121774 Make prediction for 5010 samples... 0.28499928 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 722 [0/25046 (0%)] Loss: 0.107089 Train epoch: 722 [331160/25046 (41%)] Loss: 0.132785 Train epoch: 722 [658920/25046 (82%)] Loss: 0.089735 Make prediction for 5010 samples... 0.28019994 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 723 [0/25046 (0%)] Loss: 0.082452 Train epoch: 723 [330520/25046 (41%)] Loss: 0.111900 Train epoch: 723 [659960/25046 (82%)] Loss: 0.142517 Make prediction for 5010 samples... 0.31511328 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 724 [0/25046 (0%)] Loss: 0.164206 Train epoch: 724 [323600/25046 (41%)] Loss: 0.088373 Train epoch: 724 [659400/25046 (82%)] Loss: 0.090953 Make prediction for 5010 samples... 0.27439785 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 725 [0/25046 (0%)] Loss: 0.102220 Train epoch: 725 [321260/25046 (41%)] Loss: 0.110742 Train epoch: 725 [654960/25046 (82%)] Loss: 0.083383 Make prediction for 5010 samples... 0.2758756 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 726 [0/25046 (0%)] Loss: 0.093462 Train epoch: 726 [338920/25046 (41%)] Loss: 0.115434 Train epoch: 726 [653720/25046 (82%)] Loss: 0.108410 Make prediction for 5010 samples... 0.28037295 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 727 [0/25046 (0%)] Loss: 0.107336 Train epoch: 727 [322440/25046 (41%)] Loss: 0.079617 Train epoch: 727 [664000/25046 (82%)] Loss: 0.102791 Make prediction for 5010 samples... 0.27582496 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 728 [0/25046 (0%)] Loss: 0.121150 Train epoch: 728 [327420/25046 (41%)] Loss: 0.095821 Train epoch: 728 [651800/25046 (82%)] Loss: 0.127366 Make prediction for 5010 samples... 0.29833385 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 729 [0/25046 (0%)] Loss: 0.083283 Train epoch: 729 [326140/25046 (41%)] Loss: 0.127795 Train epoch: 729 [659200/25046 (82%)] Loss: 0.140182 Make prediction for 5010 samples... 0.3014656 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 730 [0/25046 (0%)] Loss: 0.099607 Train epoch: 730 [327340/25046 (41%)] Loss: 0.088342 Train epoch: 730 [657600/25046 (82%)] Loss: 0.153002 Make prediction for 5010 samples... 0.2791135 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 731 [0/25046 (0%)] Loss: 0.104899 Train epoch: 731 [328220/25046 (41%)] Loss: 0.103436 Train epoch: 731 [648120/25046 (82%)] Loss: 0.114138 Make prediction for 5010 samples... 0.27625147 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 732 [0/25046 (0%)] Loss: 0.083002 Train epoch: 732 [329360/25046 (41%)] Loss: 0.114261 Train epoch: 732 [652920/25046 (82%)] Loss: 0.101967 Make prediction for 5010 samples... 0.29064435 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 733 [0/25046 (0%)] Loss: 0.080040 Train epoch: 733 [329460/25046 (41%)] Loss: 0.108175 Train epoch: 733 [655080/25046 (82%)] Loss: 0.099721 Make prediction for 5010 samples... 0.27162585 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 734 [0/25046 (0%)] Loss: 0.116551 Train epoch: 734 [328100/25046 (41%)] Loss: 0.094428 Train epoch: 734 [652360/25046 (82%)] Loss: 0.136573 Make prediction for 5010 samples... 0.2793369 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 735 [0/25046 (0%)] Loss: 0.105830 Train epoch: 735 [325900/25046 (41%)] Loss: 0.109153 Train epoch: 735 [666200/25046 (82%)] Loss: 0.128652 Make prediction for 5010 samples... 0.2762442 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 736 [0/25046 (0%)] Loss: 0.089381 Train epoch: 736 [331560/25046 (41%)] Loss: 0.126643 Train epoch: 736 [646720/25046 (82%)] Loss: 0.124437 Make prediction for 5010 samples... 0.28724775 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 737 [0/25046 (0%)] Loss: 0.078269 Train epoch: 737 [328140/25046 (41%)] Loss: 0.087222 Train epoch: 737 [660200/25046 (82%)] Loss: 0.094628 Make prediction for 5010 samples... 0.32149622 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 738 [0/25046 (0%)] Loss: 0.127379 Train epoch: 738 [329300/25046 (41%)] Loss: 0.098440 Train epoch: 738 [654480/25046 (82%)] Loss: 0.111175 Make prediction for 5010 samples... 0.27408904 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 739 [0/25046 (0%)] Loss: 0.126243 Train epoch: 739 [327920/25046 (41%)] Loss: 0.113601 Train epoch: 739 [655200/25046 (82%)] Loss: 0.079742 Make prediction for 5010 samples... 0.2866531 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 740 [0/25046 (0%)] Loss: 0.083242 Train epoch: 740 [331020/25046 (41%)] Loss: 0.145465 Train epoch: 740 [657240/25046 (82%)] Loss: 0.085267 Make prediction for 5010 samples... 0.27669016 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 741 [0/25046 (0%)] Loss: 0.098013 Train epoch: 741 [331380/25046 (41%)] Loss: 0.112015 Train epoch: 741 [655000/25046 (82%)] Loss: 0.102872 Make prediction for 5010 samples... 0.30679333 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 742 [0/25046 (0%)] Loss: 0.090989 Train epoch: 742 [331160/25046 (41%)] Loss: 0.108996 Train epoch: 742 [654720/25046 (82%)] Loss: 0.134035 Make prediction for 5010 samples... 0.2853843 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 743 [0/25046 (0%)] Loss: 0.124496 Train epoch: 743 [328300/25046 (41%)] Loss: 0.099004 Train epoch: 743 [651480/25046 (82%)] Loss: 0.128414 Make prediction for 5010 samples... 0.3217269 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 744 [0/25046 (0%)] Loss: 0.123862 Train epoch: 744 [324320/25046 (41%)] Loss: 0.110498 Train epoch: 744 [650080/25046 (82%)] Loss: 0.104420 Make prediction for 5010 samples... 0.28565338 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 745 [0/25046 (0%)] Loss: 0.085119 Train epoch: 745 [324660/25046 (41%)] Loss: 0.106324 Train epoch: 745 [656480/25046 (82%)] Loss: 0.112955 Make prediction for 5010 samples... 0.2708225 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 746 [0/25046 (0%)] Loss: 0.101813 Train epoch: 746 [335600/25046 (41%)] Loss: 0.087059 Train epoch: 746 [644600/25046 (82%)] Loss: 0.093800 Make prediction for 5010 samples... 0.27311826 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 747 [0/25046 (0%)] Loss: 0.082155 Train epoch: 747 [329940/25046 (41%)] Loss: 0.094165 Train epoch: 747 [660080/25046 (82%)] Loss: 0.101821 Make prediction for 5010 samples... 0.27183717 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 748 [0/25046 (0%)] Loss: 0.116119 Train epoch: 748 [326580/25046 (41%)] Loss: 0.100528 Train epoch: 748 [658240/25046 (82%)] Loss: 0.102946 Make prediction for 5010 samples... 0.268292 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 749 [0/25046 (0%)] Loss: 0.135391 Train epoch: 749 [331120/25046 (41%)] Loss: 0.124594 Train epoch: 749 [656880/25046 (82%)] Loss: 0.108667 Make prediction for 5010 samples... 0.2942671 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 750 [0/25046 (0%)] Loss: 0.087375 Train epoch: 750 [322860/25046 (41%)] Loss: 0.112918 Train epoch: 750 [659120/25046 (82%)] Loss: 0.084520 Make prediction for 5010 samples... 0.3139224 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 751 [0/25046 (0%)] Loss: 0.092108 Train epoch: 751 [328060/25046 (41%)] Loss: 0.086302 Train epoch: 751 [663400/25046 (82%)] Loss: 0.126336 Make prediction for 5010 samples... 0.29441345 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 752 [0/25046 (0%)] Loss: 0.107128 Train epoch: 752 [330580/25046 (41%)] Loss: 0.102549 Train epoch: 752 [661040/25046 (82%)] Loss: 0.120046 Make prediction for 5010 samples... 0.2998281 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 753 [0/25046 (0%)] Loss: 0.071128 Train epoch: 753 [330780/25046 (41%)] Loss: 0.122335 Train epoch: 753 [657800/25046 (82%)] Loss: 0.096590 Make prediction for 5010 samples... 0.31072405 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 754 [0/25046 (0%)] Loss: 0.118859 Train epoch: 754 [329180/25046 (41%)] Loss: 0.124021 Train epoch: 754 [663600/25046 (82%)] Loss: 0.123120 Make prediction for 5010 samples... 0.29366326 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 755 [0/25046 (0%)] Loss: 0.138150 Train epoch: 755 [327860/25046 (41%)] Loss: 0.097612 Train epoch: 755 [661280/25046 (82%)] Loss: 0.086787 Make prediction for 5010 samples... 0.27503476 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 756 [0/25046 (0%)] Loss: 0.099401 Train epoch: 756 [326100/25046 (41%)] Loss: 0.085900 Train epoch: 756 [656280/25046 (82%)] Loss: 0.119891 Make prediction for 5010 samples... 0.33841047 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 757 [0/25046 (0%)] Loss: 0.109014 Train epoch: 757 [328560/25046 (41%)] Loss: 0.071815 Train epoch: 757 [659040/25046 (82%)] Loss: 0.097509 Make prediction for 5010 samples... 0.28810933 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 758 [0/25046 (0%)] Loss: 0.091833 Train epoch: 758 [333680/25046 (41%)] Loss: 0.119334 Train epoch: 758 [651640/25046 (82%)] Loss: 0.116923 Make prediction for 5010 samples... 0.279897 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 759 [0/25046 (0%)] Loss: 0.091807 Train epoch: 759 [325720/25046 (41%)] Loss: 0.103326 Train epoch: 759 [650000/25046 (82%)] Loss: 0.109719 Make prediction for 5010 samples... 0.27634364 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 760 [0/25046 (0%)] Loss: 0.116634 Train epoch: 760 [328480/25046 (41%)] Loss: 0.121667 Train epoch: 760 [659040/25046 (82%)] Loss: 0.111462 Make prediction for 5010 samples... 0.3207328 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 761 [0/25046 (0%)] Loss: 0.136970 Train epoch: 761 [328100/25046 (41%)] Loss: 0.118716 Train epoch: 761 [664200/25046 (82%)] Loss: 0.117081 Make prediction for 5010 samples... 0.27516812 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 762 [0/25046 (0%)] Loss: 0.131362 Train epoch: 762 [324520/25046 (41%)] Loss: 0.086021 Train epoch: 762 [648320/25046 (82%)] Loss: 0.128800 Make prediction for 5010 samples... 0.2741351 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 763 [0/25046 (0%)] Loss: 0.093558 Train epoch: 763 [327880/25046 (41%)] Loss: 0.116163 Train epoch: 763 [644920/25046 (82%)] Loss: 0.118538 Make prediction for 5010 samples... 0.27420643 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 764 [0/25046 (0%)] Loss: 0.067575 Train epoch: 764 [331460/25046 (41%)] Loss: 0.103161 Train epoch: 764 [654240/25046 (82%)] Loss: 0.078626 Make prediction for 5010 samples... 0.30228776 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 765 [0/25046 (0%)] Loss: 0.078269 Train epoch: 765 [326960/25046 (41%)] Loss: 0.083527 Train epoch: 765 [658880/25046 (82%)] Loss: 0.112005 Make prediction for 5010 samples... 0.30109715 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 766 [0/25046 (0%)] Loss: 0.113553 Train epoch: 766 [331520/25046 (41%)] Loss: 0.141239 Train epoch: 766 [663440/25046 (82%)] Loss: 0.110526 Make prediction for 5010 samples... 0.2791766 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 767 [0/25046 (0%)] Loss: 0.078430 Train epoch: 767 [333880/25046 (41%)] Loss: 0.116226 Train epoch: 767 [658960/25046 (82%)] Loss: 0.088878 Make prediction for 5010 samples... 0.2841598 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 768 [0/25046 (0%)] Loss: 0.121953 Train epoch: 768 [326540/25046 (41%)] Loss: 0.164510 Train epoch: 768 [653880/25046 (82%)] Loss: 0.110626 Make prediction for 5010 samples... 0.27263895 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 769 [0/25046 (0%)] Loss: 0.099790 Train epoch: 769 [325940/25046 (41%)] Loss: 0.100218 Train epoch: 769 [659880/25046 (82%)] Loss: 0.107597 Make prediction for 5010 samples... 0.28873464 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 770 [0/25046 (0%)] Loss: 0.111461 Train epoch: 770 [326900/25046 (41%)] Loss: 0.091028 Train epoch: 770 [659120/25046 (82%)] Loss: 0.094771 Make prediction for 5010 samples... 0.28070334 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 771 [0/25046 (0%)] Loss: 0.101776 Train epoch: 771 [326420/25046 (41%)] Loss: 0.115537 Train epoch: 771 [651840/25046 (82%)] Loss: 0.109996 Make prediction for 5010 samples... 0.2753937 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 772 [0/25046 (0%)] Loss: 0.115429 Train epoch: 772 [324020/25046 (41%)] Loss: 0.097683 Train epoch: 772 [661360/25046 (82%)] Loss: 0.079178 Make prediction for 5010 samples... 0.27986628 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 773 [0/25046 (0%)] Loss: 0.107713 Train epoch: 773 [328660/25046 (41%)] Loss: 0.104051 Train epoch: 773 [648960/25046 (82%)] Loss: 0.095187 Make prediction for 5010 samples... 0.27050576 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 774 [0/25046 (0%)] Loss: 0.090885 Train epoch: 774 [325260/25046 (41%)] Loss: 0.098935 Train epoch: 774 [659800/25046 (82%)] Loss: 0.148582 Make prediction for 5010 samples... 0.28579736 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 775 [0/25046 (0%)] Loss: 0.138937 Train epoch: 775 [328000/25046 (41%)] Loss: 0.076717 Train epoch: 775 [648600/25046 (82%)] Loss: 0.065706 Make prediction for 5010 samples... 0.28275228 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 776 [0/25046 (0%)] Loss: 0.069476 Train epoch: 776 [327660/25046 (41%)] Loss: 0.141151 Train epoch: 776 [661560/25046 (82%)] Loss: 0.092424 Make prediction for 5010 samples... 0.2986539 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 777 [0/25046 (0%)] Loss: 0.095415 Train epoch: 777 [324600/25046 (41%)] Loss: 0.092217 Train epoch: 777 [661120/25046 (82%)] Loss: 0.144029 Make prediction for 5010 samples... 0.30009043 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 778 [0/25046 (0%)] Loss: 0.110910 Train epoch: 778 [325260/25046 (41%)] Loss: 0.092932 Train epoch: 778 [653960/25046 (82%)] Loss: 0.110832 Make prediction for 5010 samples... 0.2644204 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 779 [0/25046 (0%)] Loss: 0.097553 Train epoch: 779 [329820/25046 (41%)] Loss: 0.126043 Train epoch: 779 [664920/25046 (82%)] Loss: 0.094393 Make prediction for 5010 samples... 0.29138574 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 780 [0/25046 (0%)] Loss: 0.109202 Train epoch: 780 [327040/25046 (41%)] Loss: 0.107613 Train epoch: 780 [654480/25046 (82%)] Loss: 0.124738 Make prediction for 5010 samples... 0.26706797 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 781 [0/25046 (0%)] Loss: 0.096659 Train epoch: 781 [325260/25046 (41%)] Loss: 0.086978 Train epoch: 781 [655120/25046 (82%)] Loss: 0.092442 Make prediction for 5010 samples... 0.26641917 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 782 [0/25046 (0%)] Loss: 0.073001 Train epoch: 782 [326480/25046 (41%)] Loss: 0.095820 Train epoch: 782 [660520/25046 (82%)] Loss: 0.106866 Make prediction for 5010 samples... 0.3052426 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 783 [0/25046 (0%)] Loss: 0.126152 Train epoch: 783 [326260/25046 (41%)] Loss: 0.117981 Train epoch: 783 [665320/25046 (82%)] Loss: 0.106855 Make prediction for 5010 samples... 0.28601494 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 784 [0/25046 (0%)] Loss: 0.090349 Train epoch: 784 [332480/25046 (41%)] Loss: 0.091108 Train epoch: 784 [664560/25046 (82%)] Loss: 0.091707 Make prediction for 5010 samples... 0.28731552 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 785 [0/25046 (0%)] Loss: 0.100028 Train epoch: 785 [327980/25046 (41%)] Loss: 0.095982 Train epoch: 785 [654280/25046 (82%)] Loss: 0.119785 Make prediction for 5010 samples... 0.2801147 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 786 [0/25046 (0%)] Loss: 0.128598 Train epoch: 786 [328680/25046 (41%)] Loss: 0.091691 Train epoch: 786 [657000/25046 (82%)] Loss: 0.100218 Make prediction for 5010 samples... 0.27439082 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 787 [0/25046 (0%)] Loss: 0.077560 Train epoch: 787 [324860/25046 (41%)] Loss: 0.079746 Train epoch: 787 [651520/25046 (82%)] Loss: 0.083628 Make prediction for 5010 samples... 0.27848127 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 788 [0/25046 (0%)] Loss: 0.073911 Train epoch: 788 [329700/25046 (41%)] Loss: 0.091159 Train epoch: 788 [659000/25046 (82%)] Loss: 0.121241 Make prediction for 5010 samples... 0.2757991 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 789 [0/25046 (0%)] Loss: 0.096368 Train epoch: 789 [326640/25046 (41%)] Loss: 0.096644 Train epoch: 789 [661240/25046 (82%)] Loss: 0.112709 Make prediction for 5010 samples... 0.27238837 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 790 [0/25046 (0%)] Loss: 0.099681 Train epoch: 790 [328680/25046 (41%)] Loss: 0.102415 Train epoch: 790 [656040/25046 (82%)] Loss: 0.070655 Make prediction for 5010 samples... 0.2715241 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 791 [0/25046 (0%)] Loss: 0.079353 Train epoch: 791 [331300/25046 (41%)] Loss: 0.076671 Train epoch: 791 [665440/25046 (82%)] Loss: 0.082210 Make prediction for 5010 samples... 0.29059646 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 792 [0/25046 (0%)] Loss: 0.079352 Train epoch: 792 [327260/25046 (41%)] Loss: 0.084878 Train epoch: 792 [657360/25046 (82%)] Loss: 0.106363 Make prediction for 5010 samples... 0.293303 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 793 [0/25046 (0%)] Loss: 0.094136 Train epoch: 793 [325740/25046 (41%)] Loss: 0.125025 Train epoch: 793 [658480/25046 (82%)] Loss: 0.090327 Make prediction for 5010 samples... 0.27597755 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 794 [0/25046 (0%)] Loss: 0.115448 Train epoch: 794 [330940/25046 (41%)] Loss: 0.090577 Train epoch: 794 [648080/25046 (82%)] Loss: 0.134018 Make prediction for 5010 samples... 0.2772526 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 795 [0/25046 (0%)] Loss: 0.082674 Train epoch: 795 [327400/25046 (41%)] Loss: 0.109921 Train epoch: 795 [651760/25046 (82%)] Loss: 0.089643 Make prediction for 5010 samples... 0.2709179 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 796 [0/25046 (0%)] Loss: 0.080524 Train epoch: 796 [334760/25046 (41%)] Loss: 0.087666 Train epoch: 796 [655640/25046 (82%)] Loss: 0.092702 Make prediction for 5010 samples... 0.26886728 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 797 [0/25046 (0%)] Loss: 0.082243 Train epoch: 797 [332300/25046 (41%)] Loss: 0.086268 Train epoch: 797 [644080/25046 (82%)] Loss: 0.103228 Make prediction for 5010 samples... 0.2814978 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 798 [0/25046 (0%)] Loss: 0.110471 Train epoch: 798 [329880/25046 (41%)] Loss: 0.100516 Train epoch: 798 [653920/25046 (82%)] Loss: 0.076261 Make prediction for 5010 samples... 0.27138346 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 799 [0/25046 (0%)] Loss: 0.112353 Train epoch: 799 [325740/25046 (41%)] Loss: 0.110635 Train epoch: 799 [659280/25046 (82%)] Loss: 0.094691 Make prediction for 5010 samples... 0.28090876 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 800 [0/25046 (0%)] Loss: 0.069972 Train epoch: 800 [326600/25046 (41%)] Loss: 0.088200 Train epoch: 800 [657840/25046 (82%)] Loss: 0.090967 Make prediction for 5010 samples... 0.27366182 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 801 [0/25046 (0%)] Loss: 0.106349 Train epoch: 801 [327940/25046 (41%)] Loss: 0.084108 Train epoch: 801 [662360/25046 (82%)] Loss: 0.088045 Make prediction for 5010 samples... 0.28749207 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 802 [0/25046 (0%)] Loss: 0.092537 Train epoch: 802 [328400/25046 (41%)] Loss: 0.102726 Train epoch: 802 [653480/25046 (82%)] Loss: 0.101287 Make prediction for 5010 samples... 0.2930092 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 803 [0/25046 (0%)] Loss: 0.081598 Train epoch: 803 [326660/25046 (41%)] Loss: 0.092177 Train epoch: 803 [654000/25046 (82%)] Loss: 0.087462 Make prediction for 5010 samples... 0.27059773 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 804 [0/25046 (0%)] Loss: 0.093451 Train epoch: 804 [334360/25046 (41%)] Loss: 0.097293 Train epoch: 804 [656800/25046 (82%)] Loss: 0.067280 Make prediction for 5010 samples... 0.3215683 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 805 [0/25046 (0%)] Loss: 0.102464 Train epoch: 805 [327960/25046 (41%)] Loss: 0.121028 Train epoch: 805 [661480/25046 (82%)] Loss: 0.107601 Make prediction for 5010 samples... 0.2796584 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 806 [0/25046 (0%)] Loss: 0.081431 Train epoch: 806 [324080/25046 (41%)] Loss: 0.087454 Train epoch: 806 [661160/25046 (82%)] Loss: 0.146020 Make prediction for 5010 samples... 0.3011654 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 807 [0/25046 (0%)] Loss: 0.091148 Train epoch: 807 [329720/25046 (41%)] Loss: 0.121714 Train epoch: 807 [653160/25046 (82%)] Loss: 0.102569 Make prediction for 5010 samples... 0.28952414 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 808 [0/25046 (0%)] Loss: 0.101572 Train epoch: 808 [326700/25046 (41%)] Loss: 0.089781 Train epoch: 808 [655040/25046 (82%)] Loss: 0.124431 Make prediction for 5010 samples... 0.28890046 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 809 [0/25046 (0%)] Loss: 0.088085 Train epoch: 809 [331340/25046 (41%)] Loss: 0.093679 Train epoch: 809 [664080/25046 (82%)] Loss: 0.110585 Make prediction for 5010 samples... 0.29103345 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 810 [0/25046 (0%)] Loss: 0.091769 Train epoch: 810 [331300/25046 (41%)] Loss: 0.120831 Train epoch: 810 [663200/25046 (82%)] Loss: 0.153122 Make prediction for 5010 samples... 0.31649005 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 811 [0/25046 (0%)] Loss: 0.105300 Train epoch: 811 [326240/25046 (41%)] Loss: 0.091437 Train epoch: 811 [661000/25046 (82%)] Loss: 0.078282 Make prediction for 5010 samples... 0.29057202 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 812 [0/25046 (0%)] Loss: 0.110123 Train epoch: 812 [328720/25046 (41%)] Loss: 0.072804 Train epoch: 812 [659960/25046 (82%)] Loss: 0.148346 Make prediction for 5010 samples... 0.28606036 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 813 [0/25046 (0%)] Loss: 0.085519 Train epoch: 813 [330700/25046 (41%)] Loss: 0.110282 Train epoch: 813 [658840/25046 (82%)] Loss: 0.098551 Make prediction for 5010 samples... 0.27685955 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 814 [0/25046 (0%)] Loss: 0.094524 Train epoch: 814 [331300/25046 (41%)] Loss: 0.112640 Train epoch: 814 [649800/25046 (82%)] Loss: 0.082910 Make prediction for 5010 samples... 0.28782195 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 815 [0/25046 (0%)] Loss: 0.149392 Train epoch: 815 [324680/25046 (41%)] Loss: 0.080205 Train epoch: 815 [661800/25046 (82%)] Loss: 0.168938 Make prediction for 5010 samples... 0.28090504 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 816 [0/25046 (0%)] Loss: 0.077244 Train epoch: 816 [327620/25046 (41%)] Loss: 0.117301 Train epoch: 816 [654200/25046 (82%)] Loss: 0.092041 Make prediction for 5010 samples... 0.26590377 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 817 [0/25046 (0%)] Loss: 0.066605 Train epoch: 817 [331940/25046 (41%)] Loss: 0.101893 Train epoch: 817 [660640/25046 (82%)] Loss: 0.096858 Make prediction for 5010 samples... 0.27688393 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 818 [0/25046 (0%)] Loss: 0.107317 Train epoch: 818 [329140/25046 (41%)] Loss: 0.092078 Train epoch: 818 [653280/25046 (82%)] Loss: 0.117704 Make prediction for 5010 samples... 0.26983622 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 819 [0/25046 (0%)] Loss: 0.107148 Train epoch: 819 [328880/25046 (41%)] Loss: 0.071796 Train epoch: 819 [662360/25046 (82%)] Loss: 0.108995 Make prediction for 5010 samples... 0.2745167 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 820 [0/25046 (0%)] Loss: 0.081597 Train epoch: 820 [325340/25046 (41%)] Loss: 0.120652 Train epoch: 820 [657880/25046 (82%)] Loss: 0.136609 Make prediction for 5010 samples... 0.27835512 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 821 [0/25046 (0%)] Loss: 0.121136 Train epoch: 821 [331900/25046 (41%)] Loss: 0.078124 Train epoch: 821 [647680/25046 (82%)] Loss: 0.086345 Make prediction for 5010 samples... 0.27514577 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 822 [0/25046 (0%)] Loss: 0.069708 Train epoch: 822 [331240/25046 (41%)] Loss: 0.095288 Train epoch: 822 [658520/25046 (82%)] Loss: 0.105947 Make prediction for 5010 samples... 0.29043585 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 823 [0/25046 (0%)] Loss: 0.074612 Train epoch: 823 [328700/25046 (41%)] Loss: 0.111416 Train epoch: 823 [652480/25046 (82%)] Loss: 0.103559 Make prediction for 5010 samples... 0.27386144 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 824 [0/25046 (0%)] Loss: 0.090227 Train epoch: 824 [325860/25046 (41%)] Loss: 0.088562 Train epoch: 824 [657360/25046 (82%)] Loss: 0.082447 Make prediction for 5010 samples... 0.2803567 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 825 [0/25046 (0%)] Loss: 0.103547 Train epoch: 825 [328120/25046 (41%)] Loss: 0.083014 Train epoch: 825 [654800/25046 (82%)] Loss: 0.128902 Make prediction for 5010 samples... 0.28150398 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 826 [0/25046 (0%)] Loss: 0.094859 Train epoch: 826 [327980/25046 (41%)] Loss: 0.079262 Train epoch: 826 [660520/25046 (82%)] Loss: 0.106548 Make prediction for 5010 samples... 0.27289265 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 827 [0/25046 (0%)] Loss: 0.089784 Train epoch: 827 [325700/25046 (41%)] Loss: 0.095865 Train epoch: 827 [656400/25046 (82%)] Loss: 0.120453 Make prediction for 5010 samples... 0.30147746 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 828 [0/25046 (0%)] Loss: 0.090353 Train epoch: 828 [334320/25046 (41%)] Loss: 0.086907 Train epoch: 828 [657760/25046 (82%)] Loss: 0.096397 Make prediction for 5010 samples... 0.29033676 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 829 [0/25046 (0%)] Loss: 0.104635 Train epoch: 829 [331000/25046 (41%)] Loss: 0.116103 Train epoch: 829 [653400/25046 (82%)] Loss: 0.116493 Make prediction for 5010 samples... 0.27832395 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 830 [0/25046 (0%)] Loss: 0.089540 Train epoch: 830 [333060/25046 (41%)] Loss: 0.085448 Train epoch: 830 [651880/25046 (82%)] Loss: 0.108744 Make prediction for 5010 samples... 0.34072447 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 831 [0/25046 (0%)] Loss: 0.119510 Train epoch: 831 [326400/25046 (41%)] Loss: 0.089050 Train epoch: 831 [660960/25046 (82%)] Loss: 0.096063 Make prediction for 5010 samples... 0.30531326 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 832 [0/25046 (0%)] Loss: 0.116422 Train epoch: 832 [326000/25046 (41%)] Loss: 0.091617 Train epoch: 832 [660560/25046 (82%)] Loss: 0.138346 Make prediction for 5010 samples... 0.28075528 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 833 [0/25046 (0%)] Loss: 0.112320 Train epoch: 833 [327760/25046 (41%)] Loss: 0.087320 Train epoch: 833 [656840/25046 (82%)] Loss: 0.075583 Make prediction for 5010 samples... 0.2822009 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 834 [0/25046 (0%)] Loss: 0.077066 Train epoch: 834 [326020/25046 (41%)] Loss: 0.098370 Train epoch: 834 [650720/25046 (82%)] Loss: 0.095038 Make prediction for 5010 samples... 0.27571535 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 835 [0/25046 (0%)] Loss: 0.083656 Train epoch: 835 [327240/25046 (41%)] Loss: 0.098884 Train epoch: 835 [666200/25046 (82%)] Loss: 0.085199 Make prediction for 5010 samples... 0.30451426 No improvement since epoch 674 ; best_mse,best_ci: 0.26416853 0.8733403803975016 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 836 [0/25046 (0%)] Loss: 0.115027 Train epoch: 836 [328560/25046 (41%)] Loss: 0.077345 Train epoch: 836 [658840/25046 (82%)] Loss: 0.103264 Make prediction for 5010 samples... rmse improved at epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 837 [0/25046 (0%)] Loss: 0.094409 Train epoch: 837 [323540/25046 (41%)] Loss: 0.114276 Train epoch: 837 [655080/25046 (82%)] Loss: 0.128164 Make prediction for 5010 samples... 0.29449233 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 838 [0/25046 (0%)] Loss: 0.129521 Train epoch: 838 [327780/25046 (41%)] Loss: 0.106287 Train epoch: 838 [655240/25046 (82%)] Loss: 0.102410 Make prediction for 5010 samples... 0.27512246 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 839 [0/25046 (0%)] Loss: 0.084271 Train epoch: 839 [322020/25046 (41%)] Loss: 0.144296 Train epoch: 839 [653160/25046 (82%)] Loss: 0.081784 Make prediction for 5010 samples... 0.28353837 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 840 [0/25046 (0%)] Loss: 0.107776 Train epoch: 840 [331180/25046 (41%)] Loss: 0.093057 Train epoch: 840 [652600/25046 (82%)] Loss: 0.094489 Make prediction for 5010 samples... 0.27238643 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 841 [0/25046 (0%)] Loss: 0.064594 Train epoch: 841 [323840/25046 (41%)] Loss: 0.126031 Train epoch: 841 [656840/25046 (82%)] Loss: 0.106845 Make prediction for 5010 samples... 0.28174412 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 842 [0/25046 (0%)] Loss: 0.084049 Train epoch: 842 [325980/25046 (41%)] Loss: 0.080624 Train epoch: 842 [667360/25046 (82%)] Loss: 0.118187 Make prediction for 5010 samples... 0.27396467 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 843 [0/25046 (0%)] Loss: 0.090964 Train epoch: 843 [326420/25046 (41%)] Loss: 0.108054 Train epoch: 843 [650160/25046 (82%)] Loss: 0.087556 Make prediction for 5010 samples... 0.3012724 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 844 [0/25046 (0%)] Loss: 0.105899 Train epoch: 844 [326860/25046 (41%)] Loss: 0.090258 Train epoch: 844 [656280/25046 (82%)] Loss: 0.109247 Make prediction for 5010 samples... 0.3187908 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 845 [0/25046 (0%)] Loss: 0.109423 Train epoch: 845 [323780/25046 (41%)] Loss: 0.075041 Train epoch: 845 [659160/25046 (82%)] Loss: 0.114322 Make prediction for 5010 samples... 0.32263458 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 846 [0/25046 (0%)] Loss: 0.106572 Train epoch: 846 [325160/25046 (41%)] Loss: 0.073028 Train epoch: 846 [666080/25046 (82%)] Loss: 0.105238 Make prediction for 5010 samples... 0.27202573 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 847 [0/25046 (0%)] Loss: 0.095421 Train epoch: 847 [331240/25046 (41%)] Loss: 0.119582 Train epoch: 847 [660200/25046 (82%)] Loss: 0.091126 Make prediction for 5010 samples... 0.2717848 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 848 [0/25046 (0%)] Loss: 0.066969 Train epoch: 848 [327060/25046 (41%)] Loss: 0.072263 Train epoch: 848 [659480/25046 (82%)] Loss: 0.074826 Make prediction for 5010 samples... 0.28185004 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 849 [0/25046 (0%)] Loss: 0.110606 Train epoch: 849 [328900/25046 (41%)] Loss: 0.080775 Train epoch: 849 [652280/25046 (82%)] Loss: 0.083152 Make prediction for 5010 samples... 0.28577694 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 850 [0/25046 (0%)] Loss: 0.105725 Train epoch: 850 [326280/25046 (41%)] Loss: 0.073322 Train epoch: 850 [665520/25046 (82%)] Loss: 0.097917 Make prediction for 5010 samples... 0.2718092 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 851 [0/25046 (0%)] Loss: 0.067717 Train epoch: 851 [330120/25046 (41%)] Loss: 0.102464 Train epoch: 851 [658480/25046 (82%)] Loss: 0.102961 Make prediction for 5010 samples... 0.2749063 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 852 [0/25046 (0%)] Loss: 0.110872 Train epoch: 852 [326880/25046 (41%)] Loss: 0.107606 Train epoch: 852 [651160/25046 (82%)] Loss: 0.081969 Make prediction for 5010 samples... 0.29008102 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 853 [0/25046 (0%)] Loss: 0.109882 Train epoch: 853 [330620/25046 (41%)] Loss: 0.093637 Train epoch: 853 [660160/25046 (82%)] Loss: 0.098271 Make prediction for 5010 samples... 0.31107453 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 854 [0/25046 (0%)] Loss: 0.097624 Train epoch: 854 [327780/25046 (41%)] Loss: 0.121454 Train epoch: 854 [652760/25046 (82%)] Loss: 0.108542 Make prediction for 5010 samples... 0.2800726 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 855 [0/25046 (0%)] Loss: 0.101761 Train epoch: 855 [329580/25046 (41%)] Loss: 0.087645 Train epoch: 855 [657280/25046 (82%)] Loss: 0.096664 Make prediction for 5010 samples... 0.31304538 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 856 [0/25046 (0%)] Loss: 0.075474 Train epoch: 856 [325080/25046 (41%)] Loss: 0.083214 Train epoch: 856 [659080/25046 (82%)] Loss: 0.101077 Make prediction for 5010 samples... 0.289823 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 857 [0/25046 (0%)] Loss: 0.126013 Train epoch: 857 [328460/25046 (41%)] Loss: 0.082895 Train epoch: 857 [650640/25046 (82%)] Loss: 0.130877 Make prediction for 5010 samples... 0.28863844 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 858 [0/25046 (0%)] Loss: 0.093902 Train epoch: 858 [326460/25046 (41%)] Loss: 0.085396 Train epoch: 858 [659360/25046 (82%)] Loss: 0.091854 Make prediction for 5010 samples... 0.3035066 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 859 [0/25046 (0%)] Loss: 0.098260 Train epoch: 859 [330960/25046 (41%)] Loss: 0.104585 Train epoch: 859 [656600/25046 (82%)] Loss: 0.107730 Make prediction for 5010 samples... 0.27432454 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 860 [0/25046 (0%)] Loss: 0.091238 Train epoch: 860 [322620/25046 (41%)] Loss: 0.099058 Train epoch: 860 [654800/25046 (82%)] Loss: 0.082908 Make prediction for 5010 samples... 0.28253195 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 861 [0/25046 (0%)] Loss: 0.091362 Train epoch: 861 [326880/25046 (41%)] Loss: 0.120499 Train epoch: 861 [656920/25046 (82%)] Loss: 0.103236 Make prediction for 5010 samples... 0.27857986 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 862 [0/25046 (0%)] Loss: 0.068597 Train epoch: 862 [324580/25046 (41%)] Loss: 0.132960 Train epoch: 862 [657640/25046 (82%)] Loss: 0.103363 Make prediction for 5010 samples... 0.33873868 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 863 [0/25046 (0%)] Loss: 0.145723 Train epoch: 863 [330840/25046 (41%)] Loss: 0.072083 Train epoch: 863 [656840/25046 (82%)] Loss: 0.083873 Make prediction for 5010 samples... 0.2754972 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 864 [0/25046 (0%)] Loss: 0.112086 Train epoch: 864 [327760/25046 (41%)] Loss: 0.119621 Train epoch: 864 [648800/25046 (82%)] Loss: 0.077997 Make prediction for 5010 samples... 0.27291083 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 865 [0/25046 (0%)] Loss: 0.094027 Train epoch: 865 [328660/25046 (41%)] Loss: 0.093938 Train epoch: 865 [665600/25046 (82%)] Loss: 0.082021 Make prediction for 5010 samples... 0.27990207 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 866 [0/25046 (0%)] Loss: 0.088007 Train epoch: 866 [330900/25046 (41%)] Loss: 0.085386 Train epoch: 866 [652520/25046 (82%)] Loss: 0.100622 Make prediction for 5010 samples... 0.2758779 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 867 [0/25046 (0%)] Loss: 0.090060 Train epoch: 867 [333640/25046 (41%)] Loss: 0.088356 Train epoch: 867 [646720/25046 (82%)] Loss: 0.115681 Make prediction for 5010 samples... 0.28615263 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 868 [0/25046 (0%)] Loss: 0.072769 Train epoch: 868 [330700/25046 (41%)] Loss: 0.106952 Train epoch: 868 [657120/25046 (82%)] Loss: 0.075463 Make prediction for 5010 samples... 0.33455342 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 869 [0/25046 (0%)] Loss: 0.090369 Train epoch: 869 [331260/25046 (41%)] Loss: 0.081367 Train epoch: 869 [655480/25046 (82%)] Loss: 0.089910 Make prediction for 5010 samples... 0.26922962 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 870 [0/25046 (0%)] Loss: 0.092980 Train epoch: 870 [329280/25046 (41%)] Loss: 0.095215 Train epoch: 870 [667240/25046 (82%)] Loss: 0.082024 Make prediction for 5010 samples... 0.2907104 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 871 [0/25046 (0%)] Loss: 0.097138 Train epoch: 871 [328960/25046 (41%)] Loss: 0.118515 Train epoch: 871 [656720/25046 (82%)] Loss: 0.080111 Make prediction for 5010 samples... 0.29238585 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 872 [0/25046 (0%)] Loss: 0.088447 Train epoch: 872 [331020/25046 (41%)] Loss: 0.097616 Train epoch: 872 [653320/25046 (82%)] Loss: 0.085605 Make prediction for 5010 samples... 0.2897686 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 873 [0/25046 (0%)] Loss: 0.087970 Train epoch: 873 [331140/25046 (41%)] Loss: 0.098429 Train epoch: 873 [652200/25046 (82%)] Loss: 0.070601 Make prediction for 5010 samples... 0.27801776 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 874 [0/25046 (0%)] Loss: 0.069318 Train epoch: 874 [326720/25046 (41%)] Loss: 0.088086 Train epoch: 874 [662880/25046 (82%)] Loss: 0.113104 Make prediction for 5010 samples... 0.30263272 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 875 [0/25046 (0%)] Loss: 0.110284 Train epoch: 875 [325620/25046 (41%)] Loss: 0.086713 Train epoch: 875 [648920/25046 (82%)] Loss: 0.105383 Make prediction for 5010 samples... 0.27111062 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 876 [0/25046 (0%)] Loss: 0.107075 Train epoch: 876 [331880/25046 (41%)] Loss: 0.108837 Train epoch: 876 [651400/25046 (82%)] Loss: 0.086726 Make prediction for 5010 samples... 0.26687786 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 877 [0/25046 (0%)] Loss: 0.096726 Train epoch: 877 [328060/25046 (41%)] Loss: 0.082901 Train epoch: 877 [652480/25046 (82%)] Loss: 0.069421 Make prediction for 5010 samples... 0.27010545 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 878 [0/25046 (0%)] Loss: 0.084285 Train epoch: 878 [326900/25046 (41%)] Loss: 0.085867 Train epoch: 878 [668440/25046 (82%)] Loss: 0.088218 Make prediction for 5010 samples... 0.2961759 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 879 [0/25046 (0%)] Loss: 0.074310 Train epoch: 879 [326120/25046 (41%)] Loss: 0.082631 Train epoch: 879 [649920/25046 (82%)] Loss: 0.083750 Make prediction for 5010 samples... 0.2719653 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 880 [0/25046 (0%)] Loss: 0.114368 Train epoch: 880 [328680/25046 (41%)] Loss: 0.109858 Train epoch: 880 [667640/25046 (82%)] Loss: 0.109124 Make prediction for 5010 samples... 0.29471666 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 881 [0/25046 (0%)] Loss: 0.112912 Train epoch: 881 [323680/25046 (41%)] Loss: 0.100688 Train epoch: 881 [657840/25046 (82%)] Loss: 0.089365 Make prediction for 5010 samples... 0.2990874 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 882 [0/25046 (0%)] Loss: 0.090196 Train epoch: 882 [330880/25046 (41%)] Loss: 0.124509 Train epoch: 882 [656960/25046 (82%)] Loss: 0.081248 Make prediction for 5010 samples... 0.27635136 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 883 [0/25046 (0%)] Loss: 0.090836 Train epoch: 883 [328460/25046 (41%)] Loss: 0.092001 Train epoch: 883 [668440/25046 (82%)] Loss: 0.096546 Make prediction for 5010 samples... 0.28204352 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 884 [0/25046 (0%)] Loss: 0.104795 Train epoch: 884 [326280/25046 (41%)] Loss: 0.094996 Train epoch: 884 [657520/25046 (82%)] Loss: 0.142261 Make prediction for 5010 samples... 0.27246198 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 885 [0/25046 (0%)] Loss: 0.072760 Train epoch: 885 [325400/25046 (41%)] Loss: 0.090046 Train epoch: 885 [664160/25046 (82%)] Loss: 0.062654 Make prediction for 5010 samples... 0.27158713 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 886 [0/25046 (0%)] Loss: 0.091188 Train epoch: 886 [329080/25046 (41%)] Loss: 0.110535 Train epoch: 886 [651600/25046 (82%)] Loss: 0.078598 Make prediction for 5010 samples... 0.27103275 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 887 [0/25046 (0%)] Loss: 0.086928 Train epoch: 887 [328200/25046 (41%)] Loss: 0.101371 Train epoch: 887 [650280/25046 (82%)] Loss: 0.077151 Make prediction for 5010 samples... 0.2705951 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 888 [0/25046 (0%)] Loss: 0.104036 Train epoch: 888 [328520/25046 (41%)] Loss: 0.121501 Train epoch: 888 [654160/25046 (82%)] Loss: 0.076520 Make prediction for 5010 samples... 0.27965108 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 889 [0/25046 (0%)] Loss: 0.072918 Train epoch: 889 [322880/25046 (41%)] Loss: 0.102054 Train epoch: 889 [667560/25046 (82%)] Loss: 0.089651 Make prediction for 5010 samples... 0.2789925 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 890 [0/25046 (0%)] Loss: 0.082683 Train epoch: 890 [326960/25046 (41%)] Loss: 0.067952 Train epoch: 890 [655720/25046 (82%)] Loss: 0.087919 Make prediction for 5010 samples... 0.30869946 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 891 [0/25046 (0%)] Loss: 0.099691 Train epoch: 891 [324300/25046 (41%)] Loss: 0.079976 Train epoch: 891 [659640/25046 (82%)] Loss: 0.140875 Make prediction for 5010 samples... 0.28791574 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 892 [0/25046 (0%)] Loss: 0.096224 Train epoch: 892 [328780/25046 (41%)] Loss: 0.065423 Train epoch: 892 [654640/25046 (82%)] Loss: 0.096605 Make prediction for 5010 samples... 0.28362358 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 893 [0/25046 (0%)] Loss: 0.142023 Train epoch: 893 [328960/25046 (41%)] Loss: 0.088030 Train epoch: 893 [660800/25046 (82%)] Loss: 0.084386 Make prediction for 5010 samples... 0.27732074 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 894 [0/25046 (0%)] Loss: 0.089950 Train epoch: 894 [326100/25046 (41%)] Loss: 0.096205 Train epoch: 894 [657040/25046 (82%)] Loss: 0.104333 Make prediction for 5010 samples... 0.299878 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 895 [0/25046 (0%)] Loss: 0.088059 Train epoch: 895 [325680/25046 (41%)] Loss: 0.086602 Train epoch: 895 [660200/25046 (82%)] Loss: 0.094314 Make prediction for 5010 samples... 0.2980845 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 896 [0/25046 (0%)] Loss: 0.086271 Train epoch: 896 [327320/25046 (41%)] Loss: 0.107205 Train epoch: 896 [660200/25046 (82%)] Loss: 0.066810 Make prediction for 5010 samples... 0.27947986 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 897 [0/25046 (0%)] Loss: 0.066363 Train epoch: 897 [326220/25046 (41%)] Loss: 0.072219 Train epoch: 897 [652080/25046 (82%)] Loss: 0.112668 Make prediction for 5010 samples... 0.27973956 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 898 [0/25046 (0%)] Loss: 0.081033 Train epoch: 898 [331180/25046 (41%)] Loss: 0.084203 Train epoch: 898 [650480/25046 (82%)] Loss: 0.070426 Make prediction for 5010 samples... 0.2787886 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 899 [0/25046 (0%)] Loss: 0.093508 Train epoch: 899 [326560/25046 (41%)] Loss: 0.090840 Train epoch: 899 [650720/25046 (82%)] Loss: 0.084970 Make prediction for 5010 samples... 0.27144572 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 900 [0/25046 (0%)] Loss: 0.082307 Train epoch: 900 [324220/25046 (41%)] Loss: 0.068195 Train epoch: 900 [675600/25046 (82%)] Loss: 0.118838 Make prediction for 5010 samples... 0.27236676 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 901 [0/25046 (0%)] Loss: 0.083632 Train epoch: 901 [330380/25046 (41%)] Loss: 0.083547 Train epoch: 901 [642560/25046 (82%)] Loss: 0.081629 Make prediction for 5010 samples... 0.27572516 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 902 [0/25046 (0%)] Loss: 0.091340 Train epoch: 902 [327560/25046 (41%)] Loss: 0.087209 Train epoch: 902 [666920/25046 (82%)] Loss: 0.115257 Make prediction for 5010 samples... 0.30626822 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 903 [0/25046 (0%)] Loss: 0.077146 Train epoch: 903 [327340/25046 (41%)] Loss: 0.077350 Train epoch: 903 [655680/25046 (82%)] Loss: 0.092190 Make prediction for 5010 samples... 0.35130447 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 904 [0/25046 (0%)] Loss: 0.125394 Train epoch: 904 [328140/25046 (41%)] Loss: 0.088725 Train epoch: 904 [655600/25046 (82%)] Loss: 0.078513 Make prediction for 5010 samples... 0.27169105 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 905 [0/25046 (0%)] Loss: 0.104647 Train epoch: 905 [328320/25046 (41%)] Loss: 0.082914 Train epoch: 905 [664000/25046 (82%)] Loss: 0.079259 Make prediction for 5010 samples... 0.3006137 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 906 [0/25046 (0%)] Loss: 0.072212 Train epoch: 906 [327600/25046 (41%)] Loss: 0.076250 Train epoch: 906 [648440/25046 (82%)] Loss: 0.084224 Make prediction for 5010 samples... 0.28910878 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 907 [0/25046 (0%)] Loss: 0.070820 Train epoch: 907 [330280/25046 (41%)] Loss: 0.068338 Train epoch: 907 [657320/25046 (82%)] Loss: 0.072798 Make prediction for 5010 samples... 0.26674503 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 908 [0/25046 (0%)] Loss: 0.087042 Train epoch: 908 [330580/25046 (41%)] Loss: 0.103976 Train epoch: 908 [663760/25046 (82%)] Loss: 0.072562 Make prediction for 5010 samples... 0.2719811 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 909 [0/25046 (0%)] Loss: 0.082973 Train epoch: 909 [325540/25046 (41%)] Loss: 0.121239 Train epoch: 909 [657000/25046 (82%)] Loss: 0.111944 Make prediction for 5010 samples... 0.2704435 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 910 [0/25046 (0%)] Loss: 0.092102 Train epoch: 910 [328880/25046 (41%)] Loss: 0.086502 Train epoch: 910 [662880/25046 (82%)] Loss: 0.117082 Make prediction for 5010 samples... 0.2736845 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 911 [0/25046 (0%)] Loss: 0.115212 Train epoch: 911 [327700/25046 (41%)] Loss: 0.077419 Train epoch: 911 [659160/25046 (82%)] Loss: 0.089955 Make prediction for 5010 samples... 0.32018262 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 912 [0/25046 (0%)] Loss: 0.100454 Train epoch: 912 [325920/25046 (41%)] Loss: 0.087241 Train epoch: 912 [651400/25046 (82%)] Loss: 0.108782 Make prediction for 5010 samples... 0.26918328 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 913 [0/25046 (0%)] Loss: 0.075633 Train epoch: 913 [327320/25046 (41%)] Loss: 0.101458 Train epoch: 913 [648960/25046 (82%)] Loss: 0.112736 Make prediction for 5010 samples... 0.2664057 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 914 [0/25046 (0%)] Loss: 0.099903 Train epoch: 914 [329740/25046 (41%)] Loss: 0.088979 Train epoch: 914 [659280/25046 (82%)] Loss: 0.092779 Make prediction for 5010 samples... 0.27300084 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 915 [0/25046 (0%)] Loss: 0.072133 Train epoch: 915 [326160/25046 (41%)] Loss: 0.085583 Train epoch: 915 [656640/25046 (82%)] Loss: 0.097923 Make prediction for 5010 samples... 0.35094392 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 916 [0/25046 (0%)] Loss: 0.124119 Train epoch: 916 [327420/25046 (41%)] Loss: 0.088271 Train epoch: 916 [645280/25046 (82%)] Loss: 0.081709 Make prediction for 5010 samples... 0.266288 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 917 [0/25046 (0%)] Loss: 0.121204 Train epoch: 917 [324820/25046 (41%)] Loss: 0.109298 Train epoch: 917 [647920/25046 (82%)] Loss: 0.104117 Make prediction for 5010 samples... 0.27386677 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 918 [0/25046 (0%)] Loss: 0.070543 Train epoch: 918 [332140/25046 (41%)] Loss: 0.101199 Train epoch: 918 [649320/25046 (82%)] Loss: 0.123864 Make prediction for 5010 samples... 0.28944027 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 919 [0/25046 (0%)] Loss: 0.081167 Train epoch: 919 [328820/25046 (41%)] Loss: 0.071635 Train epoch: 919 [661760/25046 (82%)] Loss: 0.093351 Make prediction for 5010 samples... 0.2896942 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 920 [0/25046 (0%)] Loss: 0.082053 Train epoch: 920 [325000/25046 (41%)] Loss: 0.080347 Train epoch: 920 [654160/25046 (82%)] Loss: 0.086267 Make prediction for 5010 samples... 0.3016224 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 921 [0/25046 (0%)] Loss: 0.105195 Train epoch: 921 [330040/25046 (41%)] Loss: 0.113835 Train epoch: 921 [651200/25046 (82%)] Loss: 0.072606 Make prediction for 5010 samples... 0.31931528 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 922 [0/25046 (0%)] Loss: 0.188774 Train epoch: 922 [327040/25046 (41%)] Loss: 0.094215 Train epoch: 922 [661280/25046 (82%)] Loss: 0.083731 Make prediction for 5010 samples... 0.3061456 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 923 [0/25046 (0%)] Loss: 0.095251 Train epoch: 923 [330980/25046 (41%)] Loss: 0.095405 Train epoch: 923 [657720/25046 (82%)] Loss: 0.078420 Make prediction for 5010 samples... 0.27610675 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 924 [0/25046 (0%)] Loss: 0.076544 Train epoch: 924 [331360/25046 (41%)] Loss: 0.093677 Train epoch: 924 [661800/25046 (82%)] Loss: 0.079641 Make prediction for 5010 samples... 0.32311058 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 925 [0/25046 (0%)] Loss: 0.141710 Train epoch: 925 [327060/25046 (41%)] Loss: 0.132185 Train epoch: 925 [663080/25046 (82%)] Loss: 0.121811 Make prediction for 5010 samples... 0.31166768 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 926 [0/25046 (0%)] Loss: 0.097336 Train epoch: 926 [327820/25046 (41%)] Loss: 0.103412 Train epoch: 926 [657160/25046 (82%)] Loss: 0.098476 Make prediction for 5010 samples... 0.30343127 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 927 [0/25046 (0%)] Loss: 0.096343 Train epoch: 927 [328240/25046 (41%)] Loss: 0.089825 Train epoch: 927 [657640/25046 (82%)] Loss: 0.080468 Make prediction for 5010 samples... 0.27690512 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 928 [0/25046 (0%)] Loss: 0.080889 Train epoch: 928 [329660/25046 (41%)] Loss: 0.080922 Train epoch: 928 [652960/25046 (82%)] Loss: 0.064077 Make prediction for 5010 samples... 0.29287043 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 929 [0/25046 (0%)] Loss: 0.077831 Train epoch: 929 [325580/25046 (41%)] Loss: 0.106802 Train epoch: 929 [650400/25046 (82%)] Loss: 0.096916 Make prediction for 5010 samples... 0.32814956 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 930 [0/25046 (0%)] Loss: 0.102115 Train epoch: 930 [327360/25046 (41%)] Loss: 0.074368 Train epoch: 930 [657040/25046 (82%)] Loss: 0.108438 Make prediction for 5010 samples... 0.27668002 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 931 [0/25046 (0%)] Loss: 0.084838 Train epoch: 931 [323360/25046 (41%)] Loss: 0.082771 Train epoch: 931 [661280/25046 (82%)] Loss: 0.063998 Make prediction for 5010 samples... 0.2944478 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 932 [0/25046 (0%)] Loss: 0.102787 Train epoch: 932 [330920/25046 (41%)] Loss: 0.102262 Train epoch: 932 [657400/25046 (82%)] Loss: 0.085972 Make prediction for 5010 samples... 0.2765589 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 933 [0/25046 (0%)] Loss: 0.095282 Train epoch: 933 [328000/25046 (41%)] Loss: 0.108469 Train epoch: 933 [645720/25046 (82%)] Loss: 0.078612 Make prediction for 5010 samples... 0.3097589 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 934 [0/25046 (0%)] Loss: 0.128597 Train epoch: 934 [325940/25046 (41%)] Loss: 0.067745 Train epoch: 934 [650920/25046 (82%)] Loss: 0.080385 Make prediction for 5010 samples... 0.27951473 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 935 [0/25046 (0%)] Loss: 0.066166 Train epoch: 935 [326320/25046 (41%)] Loss: 0.077650 Train epoch: 935 [659120/25046 (82%)] Loss: 0.088035 Make prediction for 5010 samples... 0.28162986 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 936 [0/25046 (0%)] Loss: 0.076363 Train epoch: 936 [331300/25046 (41%)] Loss: 0.091920 Train epoch: 936 [658720/25046 (82%)] Loss: 0.113117 Make prediction for 5010 samples... 0.275251 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 937 [0/25046 (0%)] Loss: 0.084640 Train epoch: 937 [328980/25046 (41%)] Loss: 0.094542 Train epoch: 937 [652840/25046 (82%)] Loss: 0.073121 Make prediction for 5010 samples... 0.27765977 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 938 [0/25046 (0%)] Loss: 0.074644 Train epoch: 938 [325820/25046 (41%)] Loss: 0.092262 Train epoch: 938 [668320/25046 (82%)] Loss: 0.080780 Make prediction for 5010 samples... 0.27343574 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 939 [0/25046 (0%)] Loss: 0.091585 Train epoch: 939 [326360/25046 (41%)] Loss: 0.102205 Train epoch: 939 [659800/25046 (82%)] Loss: 0.086364 Make prediction for 5010 samples... 0.28813744 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 940 [0/25046 (0%)] Loss: 0.085518 Train epoch: 940 [325840/25046 (41%)] Loss: 0.088676 Train epoch: 940 [662160/25046 (82%)] Loss: 0.058455 Make prediction for 5010 samples... 0.30733246 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 941 [0/25046 (0%)] Loss: 0.114902 Train epoch: 941 [323280/25046 (41%)] Loss: 0.089030 Train epoch: 941 [666280/25046 (82%)] Loss: 0.087690 Make prediction for 5010 samples... 0.30505285 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 942 [0/25046 (0%)] Loss: 0.089698 Train epoch: 942 [330180/25046 (41%)] Loss: 0.087211 Train epoch: 942 [665120/25046 (82%)] Loss: 0.121966 Make prediction for 5010 samples... 0.32731944 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 943 [0/25046 (0%)] Loss: 0.093280 Train epoch: 943 [329980/25046 (41%)] Loss: 0.086491 Train epoch: 943 [658560/25046 (82%)] Loss: 0.095916 Make prediction for 5010 samples... 0.28393197 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 944 [0/25046 (0%)] Loss: 0.075319 Train epoch: 944 [328360/25046 (41%)] Loss: 0.100942 Train epoch: 944 [645360/25046 (82%)] Loss: 0.095439 Make prediction for 5010 samples... 0.2658027 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 945 [0/25046 (0%)] Loss: 0.085932 Train epoch: 945 [329480/25046 (41%)] Loss: 0.097657 Train epoch: 945 [661240/25046 (82%)] Loss: 0.096981 Make prediction for 5010 samples... 0.28056836 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 946 [0/25046 (0%)] Loss: 0.098881 Train epoch: 946 [331160/25046 (41%)] Loss: 0.113022 Train epoch: 946 [659240/25046 (82%)] Loss: 0.094450 Make prediction for 5010 samples... 0.27861485 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 947 [0/25046 (0%)] Loss: 0.080083 Train epoch: 947 [328820/25046 (41%)] Loss: 0.088424 Train epoch: 947 [650440/25046 (82%)] Loss: 0.140527 Make prediction for 5010 samples... 0.28426674 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 948 [0/25046 (0%)] Loss: 0.082792 Train epoch: 948 [329260/25046 (41%)] Loss: 0.075150 Train epoch: 948 [654320/25046 (82%)] Loss: 0.088040 Make prediction for 5010 samples... 0.2991728 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 949 [0/25046 (0%)] Loss: 0.082000 Train epoch: 949 [326900/25046 (41%)] Loss: 0.090714 Train epoch: 949 [650800/25046 (82%)] Loss: 0.084443 Make prediction for 5010 samples... 0.2745286 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 950 [0/25046 (0%)] Loss: 0.091060 Train epoch: 950 [325040/25046 (41%)] Loss: 0.081562 Train epoch: 950 [646720/25046 (82%)] Loss: 0.066492 Make prediction for 5010 samples... 0.27985027 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 951 [0/25046 (0%)] Loss: 0.075396 Train epoch: 951 [330520/25046 (41%)] Loss: 0.092035 Train epoch: 951 [649920/25046 (82%)] Loss: 0.083705 Make prediction for 5010 samples... 0.27802807 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 952 [0/25046 (0%)] Loss: 0.080072 Train epoch: 952 [325400/25046 (41%)] Loss: 0.101953 Train epoch: 952 [658880/25046 (82%)] Loss: 0.092005 Make prediction for 5010 samples... 0.27410868 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 953 [0/25046 (0%)] Loss: 0.074317 Train epoch: 953 [323540/25046 (41%)] Loss: 0.093635 Train epoch: 953 [655600/25046 (82%)] Loss: 0.072082 Make prediction for 5010 samples... 0.27133805 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 954 [0/25046 (0%)] Loss: 0.068178 Train epoch: 954 [328600/25046 (41%)] Loss: 0.064342 Train epoch: 954 [661520/25046 (82%)] Loss: 0.091467 Make prediction for 5010 samples... 0.27610725 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 955 [0/25046 (0%)] Loss: 0.077859 Train epoch: 955 [328200/25046 (41%)] Loss: 0.092183 Train epoch: 955 [655720/25046 (82%)] Loss: 0.074435 Make prediction for 5010 samples... 0.27234936 No improvement since epoch 836 ; best_mse,best_ci: 0.26370248 0.8737134602652328 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 956 [0/25046 (0%)] Loss: 0.084375 Train epoch: 956 [329440/25046 (41%)] Loss: 0.091346 Train epoch: 956 [654400/25046 (82%)] Loss: 0.089605 Make prediction for 5010 samples... rmse improved at epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 957 [0/25046 (0%)] Loss: 0.081608 Train epoch: 957 [332200/25046 (41%)] Loss: 0.101441 Train epoch: 957 [655320/25046 (82%)] Loss: 0.069916 Make prediction for 5010 samples... 0.28714028 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 958 [0/25046 (0%)] Loss: 0.093083 Train epoch: 958 [329860/25046 (41%)] Loss: 0.095096 Train epoch: 958 [651600/25046 (82%)] Loss: 0.082086 Make prediction for 5010 samples... 0.27885514 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 959 [0/25046 (0%)] Loss: 0.081096 Train epoch: 959 [333720/25046 (41%)] Loss: 0.083751 Train epoch: 959 [660560/25046 (82%)] Loss: 0.082165 Make prediction for 5010 samples... 0.2772593 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 960 [0/25046 (0%)] Loss: 0.085758 Train epoch: 960 [334880/25046 (41%)] Loss: 0.100614 Train epoch: 960 [652000/25046 (82%)] Loss: 0.077972 Make prediction for 5010 samples... 0.29754665 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 961 [0/25046 (0%)] Loss: 0.096461 Train epoch: 961 [329580/25046 (41%)] Loss: 0.081788 Train epoch: 961 [659160/25046 (82%)] Loss: 0.084442 Make prediction for 5010 samples... 0.275948 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 962 [0/25046 (0%)] Loss: 0.073796 Train epoch: 962 [332520/25046 (41%)] Loss: 0.090786 Train epoch: 962 [668840/25046 (82%)] Loss: 0.076788 Make prediction for 5010 samples... 0.28234664 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 963 [0/25046 (0%)] Loss: 0.075240 Train epoch: 963 [325780/25046 (41%)] Loss: 0.117567 Train epoch: 963 [656160/25046 (82%)] Loss: 0.097415 Make prediction for 5010 samples... 0.27982408 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 964 [0/25046 (0%)] Loss: 0.085955 Train epoch: 964 [330120/25046 (41%)] Loss: 0.074364 Train epoch: 964 [653600/25046 (82%)] Loss: 0.125612 Make prediction for 5010 samples... 0.2880667 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 965 [0/25046 (0%)] Loss: 0.074795 Train epoch: 965 [330600/25046 (41%)] Loss: 0.092798 Train epoch: 965 [656080/25046 (82%)] Loss: 0.096758 Make prediction for 5010 samples... 0.2681007 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 966 [0/25046 (0%)] Loss: 0.062181 Train epoch: 966 [329180/25046 (41%)] Loss: 0.094861 Train epoch: 966 [657840/25046 (82%)] Loss: 0.068964 Make prediction for 5010 samples... 0.30539435 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 967 [0/25046 (0%)] Loss: 0.078488 Train epoch: 967 [325740/25046 (41%)] Loss: 0.060594 Train epoch: 967 [648560/25046 (82%)] Loss: 0.075344 Make prediction for 5010 samples... 0.30206227 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 968 [0/25046 (0%)] Loss: 0.089234 Train epoch: 968 [334860/25046 (41%)] Loss: 0.106631 Train epoch: 968 [659000/25046 (82%)] Loss: 0.101363 Make prediction for 5010 samples... 0.31152555 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 969 [0/25046 (0%)] Loss: 0.100914 Train epoch: 969 [329680/25046 (41%)] Loss: 0.080407 Train epoch: 969 [654240/25046 (82%)] Loss: 0.090344 Make prediction for 5010 samples... 0.26548 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 970 [0/25046 (0%)] Loss: 0.102279 Train epoch: 970 [328520/25046 (41%)] Loss: 0.096063 Train epoch: 970 [664280/25046 (82%)] Loss: 0.080602 Make prediction for 5010 samples... 0.29397777 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 971 [0/25046 (0%)] Loss: 0.097658 Train epoch: 971 [326160/25046 (41%)] Loss: 0.085581 Train epoch: 971 [664080/25046 (82%)] Loss: 0.111248 Make prediction for 5010 samples... 0.2707576 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 972 [0/25046 (0%)] Loss: 0.065851 Train epoch: 972 [325080/25046 (41%)] Loss: 0.089444 Train epoch: 972 [655880/25046 (82%)] Loss: 0.083903 Make prediction for 5010 samples... 0.27690092 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 973 [0/25046 (0%)] Loss: 0.099047 Train epoch: 973 [327080/25046 (41%)] Loss: 0.090427 Train epoch: 973 [640760/25046 (82%)] Loss: 0.071449 Make prediction for 5010 samples... 0.29650462 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 974 [0/25046 (0%)] Loss: 0.089971 Train epoch: 974 [330020/25046 (41%)] Loss: 0.076201 Train epoch: 974 [654520/25046 (82%)] Loss: 0.094170 Make prediction for 5010 samples... 0.26595804 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 975 [0/25046 (0%)] Loss: 0.085702 Train epoch: 975 [326860/25046 (41%)] Loss: 0.079561 Train epoch: 975 [661480/25046 (82%)] Loss: 0.085841 Make prediction for 5010 samples... 0.26933482 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 976 [0/25046 (0%)] Loss: 0.113510 Train epoch: 976 [328680/25046 (41%)] Loss: 0.087317 Train epoch: 976 [655240/25046 (82%)] Loss: 0.086987 Make prediction for 5010 samples... 0.2839334 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 977 [0/25046 (0%)] Loss: 0.100883 Train epoch: 977 [325580/25046 (41%)] Loss: 0.082158 Train epoch: 977 [657760/25046 (82%)] Loss: 0.096784 Make prediction for 5010 samples... 0.27738857 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 978 [0/25046 (0%)] Loss: 0.098122 Train epoch: 978 [324800/25046 (41%)] Loss: 0.102836 Train epoch: 978 [652120/25046 (82%)] Loss: 0.070019 Make prediction for 5010 samples... 0.27700967 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 979 [0/25046 (0%)] Loss: 0.077439 Train epoch: 979 [326180/25046 (41%)] Loss: 0.117328 Train epoch: 979 [653800/25046 (82%)] Loss: 0.088356 Make prediction for 5010 samples... 0.27792397 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 980 [0/25046 (0%)] Loss: 0.106857 Train epoch: 980 [325980/25046 (41%)] Loss: 0.074465 Train epoch: 980 [655880/25046 (82%)] Loss: 0.095985 Make prediction for 5010 samples... 0.2931908 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 981 [0/25046 (0%)] Loss: 0.075136 Train epoch: 981 [322660/25046 (41%)] Loss: 0.078442 Train epoch: 981 [658040/25046 (82%)] Loss: 0.074602 Make prediction for 5010 samples... 0.26522842 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 982 [0/25046 (0%)] Loss: 0.090863 Train epoch: 982 [328740/25046 (41%)] Loss: 0.067109 Train epoch: 982 [656400/25046 (82%)] Loss: 0.106276 Make prediction for 5010 samples... 0.27529106 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 983 [0/25046 (0%)] Loss: 0.069502 Train epoch: 983 [330140/25046 (41%)] Loss: 0.085630 Train epoch: 983 [655920/25046 (82%)] Loss: 0.066633 Make prediction for 5010 samples... 0.27556762 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 984 [0/25046 (0%)] Loss: 0.063550 Train epoch: 984 [330100/25046 (41%)] Loss: 0.082719 Train epoch: 984 [656800/25046 (82%)] Loss: 0.077813 Make prediction for 5010 samples... 0.27862945 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 985 [0/25046 (0%)] Loss: 0.070766 Train epoch: 985 [326060/25046 (41%)] Loss: 0.089482 Train epoch: 985 [657320/25046 (82%)] Loss: 0.082044 Make prediction for 5010 samples... 0.30465963 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 986 [0/25046 (0%)] Loss: 0.072267 Train epoch: 986 [329940/25046 (41%)] Loss: 0.157761 Train epoch: 986 [656960/25046 (82%)] Loss: 0.056016 Make prediction for 5010 samples... 0.2700437 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 987 [0/25046 (0%)] Loss: 0.073093 Train epoch: 987 [327020/25046 (41%)] Loss: 0.082661 Train epoch: 987 [654440/25046 (82%)] Loss: 0.078025 Make prediction for 5010 samples... 0.27790156 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 988 [0/25046 (0%)] Loss: 0.060061 Train epoch: 988 [331040/25046 (41%)] Loss: 0.074247 Train epoch: 988 [657480/25046 (82%)] Loss: 0.085159 Make prediction for 5010 samples... 0.2820108 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 989 [0/25046 (0%)] Loss: 0.097643 Train epoch: 989 [325140/25046 (41%)] Loss: 0.072805 Train epoch: 989 [657240/25046 (82%)] Loss: 0.081109 Make prediction for 5010 samples... 0.27145308 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 990 [0/25046 (0%)] Loss: 0.078373 Train epoch: 990 [320880/25046 (41%)] Loss: 0.105150 Train epoch: 990 [654280/25046 (82%)] Loss: 0.104458 Make prediction for 5010 samples... 0.2827403 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 991 [0/25046 (0%)] Loss: 0.078091 Train epoch: 991 [329480/25046 (41%)] Loss: 0.092001 Train epoch: 991 [645000/25046 (82%)] Loss: 0.103141 Make prediction for 5010 samples... 0.26492938 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 992 [0/25046 (0%)] Loss: 0.088578 Train epoch: 992 [328620/25046 (41%)] Loss: 0.100219 Train epoch: 992 [655560/25046 (82%)] Loss: 0.104946 Make prediction for 5010 samples... 0.2804618 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 993 [0/25046 (0%)] Loss: 0.094565 Train epoch: 993 [330440/25046 (41%)] Loss: 0.086622 Train epoch: 993 [673880/25046 (82%)] Loss: 0.103179 Make prediction for 5010 samples... 0.28431267 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 994 [0/25046 (0%)] Loss: 0.080723 Train epoch: 994 [329380/25046 (41%)] Loss: 0.069769 Train epoch: 994 [649720/25046 (82%)] Loss: 0.078095 Make prediction for 5010 samples... 0.27042052 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 995 [0/25046 (0%)] Loss: 0.109117 Train epoch: 995 [331420/25046 (41%)] Loss: 0.104446 Train epoch: 995 [654920/25046 (82%)] Loss: 0.097646 Make prediction for 5010 samples... 0.2701926 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 996 [0/25046 (0%)] Loss: 0.060960 Train epoch: 996 [327420/25046 (41%)] Loss: 0.065694 Train epoch: 996 [657840/25046 (82%)] Loss: 0.066770 Make prediction for 5010 samples... 0.2835396 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 997 [0/25046 (0%)] Loss: 0.094647 Train epoch: 997 [325060/25046 (41%)] Loss: 0.079212 Train epoch: 997 [658000/25046 (82%)] Loss: 0.078683 Make prediction for 5010 samples... 0.29310802 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 998 [0/25046 (0%)] Loss: 0.078515 Train epoch: 998 [330320/25046 (41%)] Loss: 0.073112 Train epoch: 998 [658520/25046 (82%)] Loss: 0.086440 Make prediction for 5010 samples... 0.2718394 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 999 [0/25046 (0%)] Loss: 0.082572 Train epoch: 999 [324400/25046 (41%)] Loss: 0.116811 Train epoch: 999 [658400/25046 (82%)] Loss: 0.094223 Make prediction for 5010 samples... 0.26860142 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis Training on 25046 samples... Train epoch: 1000 [0/25046 (0%)] Loss: 0.140258 Train epoch: 1000 [327940/25046 (41%)] Loss: 0.081716 Train epoch: 1000 [657840/25046 (82%)] Loss: 0.104596 Make prediction for 5010 samples... 0.29935443 No improvement since epoch 956 ; best_mse,best_ci: 0.26266533 0.8721359113417658 <class '__main__.GCNNet'> davis

6. Predict with pretrained model

사전 학습 모델을 사용하면 학습 단계를 건너뛰고 예측 단계부터 시작하여 성능 평가를 진행합니다.코드는 5번 과정의 Predict, Main과 동일합니다.

6-1. Predict

def predicting(model, device, loader):
    model.eval()
    total_preds = torch.Tensor()
    total_labels = torch.Tensor()
    print('Make prediction for {} samples...'.format(len(loader.dataset)))
    with torch.no_grad():
        for data in loader:
            data = data.to(device)
            output = model(data)
            total_preds = torch.cat((total_preds, output.cpu()), 0)
            total_labels = torch.cat((total_labels, data.y.view(-1, 1).cpu()), 0)
    return total_labels.numpy().flatten(),total_preds.numpy().flatten()
Shift+Enter to run
Python
→

6-2. Sample Test

datasets = ['davis','kiba']
modelings = [GINConvNet, GATNet, GAT_GCN, GCNNet]
cuda_name = "cuda:0"
Shift+Enter to run
Python
→
TEST_BATCH_SIZE = 512
Shift+Enter to run
Python
→
print('-----------------------------------------------------')
print('Select Dataset Number...')
print('|  0 : davis  |  1 : kiba  |')
d = int(input())
print('The ' + datasets[d] + ' dataset has been selected!')
print('-----------------------------------------------------')
print('Select Model Number...')
print('|  0 : GIN  |  1 : GAT  |  2 : GAT_GCN  |  3 : GCN  |')
m = int(input())
print('The ' + modelings[m].__name__ + ' model has been selected!')
print('-----------------------------------------------------')
print('\nrunning on ', modelings[m].__name__ + '_' + datasets[d])
result = []
processed_data_file_test = '/content/drive/My Drive/GraphDTA/data/processed/' + datasets[d] + '_test.pt'
if (not os.path.isfile(processed_data_file_test)):
    print('please run create_data.py to prepare data in pytorch format!')
else:
    test_data = TestbedDataset(root='/content/drive/My Drive/GraphDTA/data', dataset=datasets[d]+'_test')
    test_loader = DataLoader(test_data, batch_size=TEST_BATCH_SIZE, shuffle=False)
    model_st = modelings[m].__name__
    print('\npredicting for ', datasets[d], ' using ', model_st)
    # training the model
    device = torch.device(cuda_name if torch.cuda.is_available() else "cpu")
    model = modelings[m]().to(device)
    model_file_name = '/content/drive/My Drive/GraphDTA/pretrained/model_' + model_st + '_' + datasets[d] +  '.model'
    if os.path.isfile(model_file_name):            
        model.load_state_dict(torch.load(model_file_name, map_location=cuda_name), strict=False)
        G,P = predicting(model, device, test_loader)
        ret = [rmse(G,P),mse(G,P),pearson(G,P),spearman(G,P),ci(G,P)]
        ret =[datasets[d], model_st] +  [round(e,3) for e in ret]
        result += [ ret ]
        print('dataset,model,rmse,mse,pearson,spearman,ci')
        print(ret)
    else:
        print('model is not available!')
Shift+Enter to run
Python
→
----------------------------------------------------- Select Dataset Number... | 0 : davis | 1 : kiba | 0 The davis dataset has been selected! ----------------------------------------------------- Select Model Number... | 0 : GIN | 1 : GAT | 2 : GAT_GCN | 3 : GCN | 0 The GINConvNet model has been selected! ----------------------------------------------------- running on GINConvNet_davis Pre-processed data found: /content/drive/My Drive/GraphDTA/data/processed/davis_test.pt, loading ... predicting for davis using GINConvNet Make prediction for 5010 samples... dataset,model,rmse,mse,pearson,spearman,ci ['davis', 'GINConvNet', 1.091, 1.189, 0.321, 0.307, 0.662] 
with open('/content/drive/My Drive/GraphDTA/sample_result_' + modelings[m].__name__ + '_' + datasets[d] + '.csv','w') as f:
    f.write('dataset,model,rmse,mse,pearson,spearman,ci\n')
    
    for ret in result:
        f.write(','.join(map(str,ret)) + '\n')
Shift+Enter to run
Python
→

6-3. Main Test

result = []
for dataset in datasets :
    processed_data_file_test = '/content/drive/My Drive/GraphDTA/data/processed/' + dataset + '_test.pt'
    if (not os.path.isfile(processed_data_file_test)):
        print('please run create_data.py to prepare data in pytorch format!')
    else:
        test_data = TestbedDataset(root='/content/drive/My Drive/GraphDTA/data', dataset=dataset+'_test')
        test_loader = DataLoader(test_data, batch_size=TEST_BATCH_SIZE, shuffle=False)
        for modeling in modelings :
            model_st = modeling.__name__
            print('\npredicting for ', dataset, ' using ', model_st)
            # training the model
            device = torch.device(cuda_name if torch.cuda.is_available() else "cpu")
            model = modeling().to(device)
            model_file_name = '/content/drive/My Drive/GraphDTA/pretrained/model_' + model_st + '_' + dataset +  '.model'
            if os.path.isfile(model_file_name):            
                model.load_state_dict(torch.load(model_file_name, map_location=cuda_name), strict=False)
                G,P = predicting(model, device, test_loader)
                ret = [rmse(G,P),mse(G,P),pearson(G,P),spearman(G,P),ci(G,P)]
                ret =[dataset, model_st] +  [round(e,3) for e in ret]
                result += [ ret ]
                print('dataset,model,rmse,mse,pearson,spearman,ci')
                print(ret)
            else:
                print('model is not available!')
    print('\n')
Shift+Enter to run
Python
→
Pre-processed data found: /content/drive/My Drive/GraphDTA/data/processed/davis_test.pt, loading ... predicting for davis using GINConvNet Make prediction for 5010 samples... dataset,model,rmse,mse,pearson,spearman,ci ['davis', 'GINConvNet', 1.091, 1.189, 0.321, 0.307, 0.662] predicting for davis using GATNet Make prediction for 5010 samples... dataset,model,rmse,mse,pearson,spearman,ci ['davis', 'GATNet', 0.981, 0.962, -0.025, -0.051, 0.473] predicting for davis using GAT_GCN Make prediction for 5010 samples... dataset,model,rmse,mse,pearson,spearman,ci ['davis', 'GAT_GCN', 1.292, 1.67, -0.042, -0.048, 0.475] predicting for davis using GCNNet Make prediction for 5010 samples... dataset,model,rmse,mse,pearson,spearman,ci ['davis', 'GCNNet', 1.129, 1.274, 0.319, 0.285, 0.651] Pre-processed data found: /content/drive/My Drive/GraphDTA/data/processed/kiba_test.pt, loading ... predicting for kiba using GINConvNet Make prediction for 19709 samples... dataset,model,rmse,mse,pearson,spearman,ci ['kiba', 'GINConvNet', 0.78, 0.608, 0.446, 0.398, 0.648] predicting for kiba using GATNet Make prediction for 19709 samples... dataset,model,rmse,mse,pearson,spearman,ci ['kiba', 'GATNet', 1.084, 1.175, 0.032, 0.067, 0.524] predicting for kiba using GAT_GCN Make prediction for 19709 samples... dataset,model,rmse,mse,pearson,spearman,ci ['kiba', 'GAT_GCN', 1.623, 2.634, -0.029, -0.045, 0.483] predicting for kiba using GCNNet Make prediction for 19709 samples... dataset,model,rmse,mse,pearson,spearman,ci ['kiba', 'GCNNet', 0.77, 0.592, 0.439, 0.377, 0.639] 
with open('/content/drive/My Drive/GraphDTA/result.csv','w') as f:
    f.write('dataset,model,rmse,mse,pearson,spearman,ci\n')
    
    for ret in result:
        f.write(','.join(map(str,ret)) + '\n')
Shift+Enter to run
Python
→

7. Result

7.1 Table

df = pd.read_csv('/content/drive/My Drive/GraphDTA/result.csv')
df
Shift+Enter to run
Python
→

7.2 Visualization

plt.rcParams['figure.figsize'] = (8,8)
Shift+Enter to run
Python
→
sns.barplot(data=df, x='dataset', y='rmse', hue='model')
Shift+Enter to run
Python
→
sns.barplot(data=df, x='dataset', y='mse', hue='model')
Shift+Enter to run
Python
→
sns.barplot(data=df, x='dataset', y='ci', hue='model')
Shift+Enter to run
Python
→
Runtimes (1)