# Draft: Keras Fashion-MNIST This notebook runs on a python environment that contains Keras, Tensorflow and Numpy. It describes a small Machine Learning challenge with a suggested solution to the challenge. Your mission is to come with a better solution to the challenge and to share it with your colleagues, customers or friends either privately or publicly. Currently, the notebook is in read-only mode. In order to change the content of a cell, all you need to do is to click on the Remix button on the right of the top bar. Once you are satisfied with your solution, share it by clicking on the Publish button that will appears once you remix the notebook. # Setup There is almost nothing to setup as the "Keras Installation" environment already contains the required python packages. The only thing we need to do is to import the required dependencies in a Python code cell: # The Challenge Fashion MNIST. ```python id=f254b56e-6d60-45a2-a74c-cd66fa9f52ae import keras, tensorflow as tf, numpy as np, matplotlib.pyplot as plt from keras.datasets import fashion_mnist np.random.seed(1) tf.set_random_seed(1) (train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data() train_images = train_images / 255.0 test_images = test_images / 255.0 class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat', 'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot'] ``` ```python id=727b7326-4a59-44ff-b2a3-90d50b67c360 fig = plt.figure(figsize=(10,10)) for i in range(25): plt.subplot(5,5,i+1) plt.xticks([]) plt.yticks([]) plt.grid(False) plt.imshow(train_images[i], cmap=plt.cm.binary) plt.xlabel(class_names[train_labels[i]]) fig ``` ![result][nextjournal#output#727b7326-4a59-44ff-b2a3-90d50b67c360#result] # The Model Define ```python id=1e24c8c3-2b42-4642-9ade-1ac31a0424f7 model = keras.Sequential([ keras.layers.Flatten(input_shape=(28, 28)), keras.layers.Dense(128, activation=tf.nn.relu), keras.layers.Dense(10, activation=tf.nn.softmax) ]) ``` Compile ```python id=3984a3e1-9bb0-4b2e-9311-f792c54f149c model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy']) ``` Now, we train our model: ```python id=281e8429-bcac-44b0-a667-83434870ecd2 model.fit(train_images, train_labels, epochs=5) ``` # Model Evaluation Let's check how our model performs: ```python id=ca46915d-3dc3-4405-aad2-dc413cdd283c test_loss, test_acc = model.evaluate(test_images, test_labels) print('Test accuracy:', test_acc) ``` Not bad! But can you do better? # Visualization ```python id=7b35c176-4fd0-4cc5-ac05-9dfddb0767e3 predictions = model.predict(test_images) ``` ```python id=197e58ea-89ee-44ff-b9ba-cb8f5aa5df99 def plot_image(i, predictions_array, true_label, img): predictions_array = predictions_array[i] true_label = true_label[i] img = img[i] plt.grid(False) plt.xticks([]) plt.yticks([]) plt.imshow(img, cmap=plt.cm.binary) predicted_label = np.argmax(predictions_array) if predicted_label == true_label: color = 'blue' else: color = 'red' plt.xlabel("{} {:2.0f}% ({})".format(class_names[predicted_label], 100*np.max(predictions_array), class_names[true_label]), color=color) ``` ```python id=2183a344-8804-404f-87df-a14fb6954d60 def plot_value_array(i, predictions_array, true_label): predictions_array, true_label = predictions_array[i], true_label[i] plt.grid(False) plt.yticks([]) plt.xticks([]) plot = plt.bar(range(10), predictions_array, color="#777777", align="center") plt.ylim([0, 1]) predicted_label = np.argmax(predictions_array) plot[predicted_label].set_color('red') plot[true_label].set_color('blue') ``` ```python id=828a5109-6a99-43e6-9167-dd5c653d15f5 i = 0 f1 = plt.figure(figsize=(10,5)) plt.subplot(1,2,1) plot_image(i, predictions, test_labels, test_images) plt.subplot(1,2,2) plot_value_array(i, predictions, test_labels) f1 ``` ![result][nextjournal#output#828a5109-6a99-43e6-9167-dd5c653d15f5#result] ```python id=54778791-70ef-47fa-8ad8-71fcafb22e31 num_rows = 5 num_cols = 3 num_images = num_rows*num_cols fig = plt.figure(figsize=(2*2*num_cols, 2*num_rows)) for i in range(num_images): plt.subplot(num_rows, 2*num_cols, 2*i+1) plot_image(i, predictions, test_labels, test_images) plt.subplot(num_rows, 2*num_cols, 2*i+2) plot_value_array(i, predictions, test_labels) fig ``` ![result][nextjournal#output#54778791-70ef-47fa-8ad8-71fcafb22e31#result] All you need to do is to click on the "Remix" button on the right of the top bar. Now the notebook becomes your own and you can author it. Once you are satisfied with your model, share it with privately with your colleagues or publicly. It is a simple as clicking on the "Publish" button. Happy notebooking! [nextjournal#output#727b7326-4a59-44ff-b2a3-90d50b67c360#result]: [nextjournal#output#828a5109-6a99-43e6-9167-dd5c653d15f5#result]: [nextjournal#output#54778791-70ef-47fa-8ad8-71fcafb22e31#result]:
This notebook was exported from https://nextjournal.com/a/KyYJvoLWF3yKNBuU35qqL?change-id=CRxZ8M65inFdLvme3bwBCE ```edn nextjournal-metadata {:article {:settings nil, :nodes {"197e58ea-89ee-44ff-b9ba-cb8f5aa5df99" {:compute-ref #uuid "0a23592f-6fd2-4014-91d4-673e5df6d6ca", :exec-duration 375, :id "197e58ea-89ee-44ff-b9ba-cb8f5aa5df99", :kind "code", :output-log-lines {:stdout 0}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "1e24c8c3-2b42-4642-9ade-1ac31a0424f7" {:compute-ref #uuid "915b21dc-cbda-49d0-907b-8f69faa90906", :exec-duration 476, :id "1e24c8c3-2b42-4642-9ade-1ac31a0424f7", :kind "code", :output-log-lines {:stdout 0}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "2183a344-8804-404f-87df-a14fb6954d60" {:compute-ref #uuid "9bfd8b75-1a54-4e69-88b8-33e399d28b71", :exec-duration 391, :id "2183a344-8804-404f-87df-a14fb6954d60", :kind "code", :output-log-lines {:stdout 0}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "281e8429-bcac-44b0-a667-83434870ecd2" {:compute-ref #uuid "9e188cd8-16ac-4bef-b29b-6edd3dd3e8c2", :exec-duration 25113, :id "281e8429-bcac-44b0-a667-83434870ecd2", :kind "code", :output-log-lines {:stdout 490}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e" {:environment [:environment {:article/nextjournal.id #uuid "59cd6481-034b-4a57-bd9b-7adc800a3546", :change/nextjournal.id #uuid "5c718fec-ff74-4e9f-8fea-5c41d1438fe2", :node/id "fbed92a0-f8ff-47e5-9ddc-aa00eae895f1"}], :id "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e", :kind "runtime", :language "python", :resources {:machine-type "n1-standard-1"}, :type :nextjournal}, "3984a3e1-9bb0-4b2e-9311-f792c54f149c" {:compute-ref #uuid "9c40b650-f9c8-4d05-90e4-76b3d5a64d7f", :exec-duration 465, :id "3984a3e1-9bb0-4b2e-9311-f792c54f149c", :kind "code", :output-log-lines {:stdout 0}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "54778791-70ef-47fa-8ad8-71fcafb22e31" {:compute-ref #uuid "6134d585-9b05-4b52-a1b3-c070c3813e51", :exec-duration 2187, :id "54778791-70ef-47fa-8ad8-71fcafb22e31", :kind "code", :output-log-lines {:stdout 0}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "727b7326-4a59-44ff-b2a3-90d50b67c360" {:compute-ref #uuid "21b71670-2100-4dea-9b54-828c3f36fa04", :exec-duration 1667, :id "727b7326-4a59-44ff-b2a3-90d50b67c360", :kind "code", :output-log-lines {:stdout 0}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "7b35c176-4fd0-4cc5-ac05-9dfddb0767e3" {:compute-ref #uuid "6eea8579-a3c7-4464-863d-4d68b4676ae2", :exec-duration 614, :id "7b35c176-4fd0-4cc5-ac05-9dfddb0767e3", :kind "code", :output-log-lines {:stdout 0}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "828a5109-6a99-43e6-9167-dd5c653d15f5" {:compute-ref #uuid "ce008e37-1891-4f53-99f8-d22950ebf664", :exec-duration 1162, :id "828a5109-6a99-43e6-9167-dd5c653d15f5", :kind "code", :output-log-lines {:stdout 0}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "ca46915d-3dc3-4405-aad2-dc413cdd283c" {:compute-ref #uuid "3568412f-fad9-485d-920a-c3b096af98db", :exec-duration 614, :id "ca46915d-3dc3-4405-aad2-dc413cdd283c", :kind "code", :output-log-lines {:stdout 9}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}, "f254b56e-6d60-45a2-a74c-cd66fa9f52ae" {:compute-ref #uuid "3f585b97-86bc-4bbb-8f64-24ecf3452ebc", :exec-duration 1500, :id "f254b56e-6d60-45a2-a74c-cd66fa9f52ae", :kind "code", :output-log-lines {:stdout 0}, :runtime [:runtime "3083bc75-7f21-4bdf-b495-5e5dc9c6af0e"]}}, :nextjournal/id #uuid "02a63b54-7012-454b-b268-d491f6459353", :article/change {:nextjournal/id #uuid "5c909caf-7219-4c41-9245-f4d70fe31a03"}}} ```