Update to Keras 2

titu1994 · May 23, 2017 · c37cde2 · c37cde2
1 parent fa600c6
commit c37cde2
Show file tree

Hide file tree

Showing 7 changed files with 107 additions and 34 deletions.
diff --git a/.idea/markdown-navigator.xml b/.idea/markdown-navigator.xml
diff --git a/.idea/markdown-navigator/profiles_settings.xml b/.idea/markdown-navigator/profiles_settings.xml
diff --git a/cifar10_wrn_16_8.py b/cifar10_wrn_16_8.py
@@ -6,7 +6,7 @@
 import keras.callbacks as callbacks
 import keras.utils.np_utils as kutils
 from keras.preprocessing.image import ImageDataGenerator
-from keras.utils.visualize_util import plot
+from keras.utils import plot_model
 
 from keras import backend as K
 
@@ -39,7 +39,7 @@
 model = wrn.create_wide_residual_network(init_shape, nb_classes=10, N=2, k=8, dropout=0.00)
 
 model.summary()
-#plot(model, "WRN-16-8.png", show_shapes=False)
+#plot_model(model, "WRN-16-8.png", show_shapes=False)
 
 model.compile(loss="categorical_crossentropy", optimizer="adam", metrics=["acc"])
 print("Finished compiling")
@@ -48,10 +48,10 @@
 print("Model loaded.")
 print("Allocating GPU memory")
 
-#model.fit_generator(generator.flow(trainX, trainY, batch_size=batch_size), samples_per_epoch=len(trainX), nb_epoch=nb_epoch,
+# model.fit_generator(generator.flow(trainX, trainY, batch_size=batch_size), steps_per_epoch=len(trainX) // batch_size + 1, nb_epoch=nb_epoch,
 #                    callbacks=[callbacks.ModelCheckpoint("WRN-16-8 Weights.h5", monitor="val_acc", save_best_only=True)],
 #                    validation_data=(testX, testY),
-#                    nb_val_samples=testX.shape[0],)
+#                    validation_steps=testX.shape[0] // batch_size + 1,)
 
 yPreds = model.predict(testX)
 yPred = np.argmax(yPreds, axis=1)

diff --git a/cifar10_wrn_28_8.py b/cifar10_wrn_28_8.py
@@ -6,7 +6,7 @@
 import keras.callbacks as callbacks
 import keras.utils.np_utils as kutils
 from keras.preprocessing.image import ImageDataGenerator
-from keras.utils.visualize_util import plot
+from keras.utils import plot_model
 
 from keras import backend as K
 
@@ -47,7 +47,7 @@
 model = wrn.create_wide_residual_network(init_shape, nb_classes=10, N=4, k=8, dropout=0.0)
 
 model.summary()
-#plot(model, "WRN-28-8.png", show_shapes=False)
+#plot_model(model, "WRN-28-8.png", show_shapes=False)
 
 model.compile(loss="categorical_crossentropy", optimizer="adadelta", metrics=["acc"])
 print("Finished compiling")
@@ -56,10 +56,10 @@
 model.load_weights("weights/WRN-28-8 Weights.h5")
 print("Model loaded.")
 
-#model.fit_generator(generator.flow(trainX, trainY, batch_size=batch_size), samples_per_epoch=len(trainX), nb_epoch=nb_epoch,
-#                    callbacks=[callbacks.ModelCheckpoint("WRN-28-8 Weights.h5", monitor="val_acc", save_best_only=True)],
-#                    validation_data=test_generator.flow(testX, testY, batch_size=batch_size),
-#                    nb_val_samples=testX.shape[0],)
+model.fit_generator(generator.flow(trainX, trainY, batch_size=batch_size), steps_per_epoch=len(trainX) // batch_size + 1, nb_epoch=nb_epoch,
+                   callbacks=[callbacks.ModelCheckpoint("WRN-28-8 Weights.h5", monitor="val_acc", save_best_only=True)],
+                   validation_data=test_generator.flow(testX, testY, batch_size=batch_size),
+                   validation_steps=testX.shape[0] // batch_size + 1,)
 
-scores = model.evaluate_generator(test_generator.flow(testX, testY, nb_epoch), testX.shape[0])
+scores = model.evaluate_generator(test_generator.flow(testX, testY, nb_epoch), testX.shape[0] // batch_size + 1)
 print("Accuracy = %f" % (100 * scores[1]))
diff --git a/weights/ADD WEIGHT FILES HERE.md b/weights/ADD WEIGHT FILES HERE.md
@@ -0,0 +1,3 @@
+Please download the approproate Weight files and extract them here.
+
+Weights for WRN-16-8 and WRN-28-8 are available [in the release tab](https://github.com/titu1994/Wide-Residual-Networks/releases)
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,3 @@
		Please download the approproate Weight files and extract them here.

		Weights for WRN-16-8 and WRN-28-8 are available [in the release tab](https://github.com/titu1994/Wide-Residual-Networks/releases)