keras update acc > accuracy

Author: bob7783

ann_class2/keras_example.py

@@ -75,8 +75,8 @@
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

ann_class2/keras_functional.py

@@ -70,8 +70,8 @@
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

cnn_class/keras_example.py

@@ -113,8 +113,8 @@ def rearrange(X):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

cnn_class2/fashion.py

@@ -101,8 +101,8 @@ def y2indicator(Y):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

cnn_class2/fashion2.py

@@ -96,8 +96,8 @@ def y2indicator(Y):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

cnn_class2/use_pretrained_weights_resnet.py

@@ -173,8 +173,8 @@ def get_confusion_matrix(data_path, N):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='train acc')
-plt.plot(r.history['val_acc'], label='val acc')
+plt.plot(r.history['accuracy'], label='train acc')
+plt.plot(r.history['val_accuracy'], label='val acc')
 plt.legend()
 plt.show()
 

cnn_class2/use_pretrained_weights_vgg.py

@@ -31,10 +31,10 @@
 # valid_path = '../large_files/blood_cell_images/TEST'
 
 # https://www.kaggle.com/moltean/fruits
-# train_path = '../large_files/fruits-360/Training'
-# valid_path = '../large_files/fruits-360/Validation'
-train_path = '../large_files/fruits-360-small/Training'
-valid_path = '../large_files/fruits-360-small/Validation'
+train_path = '../large_files/fruits-360/Training'
+valid_path = '../large_files/fruits-360/Validation'
+# train_path = '../large_files/fruits-360-small/Training'
+# valid_path = '../large_files/fruits-360-small/Validation'
 
 # useful for getting number of files
 image_files = glob(train_path + '/*/*.jp*g')
@@ -45,7 +45,7 @@
 
 
 # look at an image for fun
-plt.imshow(image.load_img(np.random.choice(image_files)))
+plt.imshow(image.img_to_array(image.load_img(np.random.choice(image_files))).astype('uint8'))
 plt.show()
 
 
@@ -76,7 +76,6 @@
 )
 
 
-
 # create an instance of ImageDataGenerator
 gen = ImageDataGenerator(
   rotation_range=20,
@@ -172,8 +171,8 @@ def get_confusion_matrix(data_path, N):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='train acc')
-plt.plot(r.history['val_acc'], label='val acc')
+plt.plot(r.history['accuracy'], label='train acc')
+plt.plot(r.history['val_accuracy'], label='val acc')
 plt.legend()
 plt.show()
 

nlp_class2/pos_ner_keras.py

@@ -217,8 +217,8 @@ def get_data_ner(split_sequences=False):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

nlp_class3/attention.py

@@ -34,10 +34,10 @@ def softmax_over_time(x):
 
 # config
 BATCH_SIZE = 64
-EPOCHS = 100
-LATENT_DIM = 256
-LATENT_DIM_DECODER = 256 # idea: make it different to ensure things all fit together properly!
-NUM_SAMPLES = 10000
+EPOCHS = 30
+LATENT_DIM = 400
+LATENT_DIM_DECODER = 400 # idea: make it different to ensure things all fit together properly!
+NUM_SAMPLES = 20000
 MAX_SEQUENCE_LENGTH = 100
 MAX_NUM_WORDS = 20000
 EMBEDDING_DIM = 100
@@ -190,7 +190,8 @@ def softmax_over_time(x):
 # assign the values
 for i, d in enumerate(decoder_targets):
   for t, word in enumerate(d):
-    decoder_targets_one_hot[i, t, word] = 1
+    if word > 0:
+      decoder_targets_one_hot[i, t, word] = 1
 
 
 
@@ -367,8 +368,8 @@ def acc(y_true, y_pred):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

nlp_class3/bilstm_mnist.py

@@ -97,8 +97,8 @@ def get_mnist(limit=None):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

nlp_class3/cnn_toxic.py

@@ -70,11 +70,15 @@
 s = sorted(len(s) for s in sequences)
 print("median sequence length:", s[len(s) // 2])
 
+print("max word index:", max(max(seq) for seq in sequences if len(seq) > 0))
+
 
 # get word -> integer mapping
 word2idx = tokenizer.word_index
 print('Found %s unique tokens.' % len(word2idx))
 
+# exit()
+
 
 # pad sequences so that we get a N x T matrix
 data = pad_sequences(sequences, maxlen=MAX_SEQUENCE_LENGTH)
@@ -144,8 +148,8 @@
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

nlp_class3/lstm_toxic.py

@@ -20,9 +20,9 @@
 from sklearn.metrics import roc_auc_score
 
 import keras.backend as K
-if len(K.tensorflow_backend._get_available_gpus()) > 0:
-  from keras.layers import CuDNNLSTM as LSTM
-  from keras.layers import CuDNNGRU as GRU
+# if len(K.tensorflow_backend._get_available_gpus()) > 0:
+#   from keras.layers import CuDNNLSTM as LSTM
+#   from keras.layers import CuDNNGRU as GRU
 
 
 # Download the data:
@@ -124,7 +124,7 @@
 model.compile(
   loss='binary_crossentropy',
   optimizer=Adam(lr=0.01),
-  metrics=['accuracy']
+  metrics=['accuracy'],
 )
 
 
@@ -144,8 +144,8 @@
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

nlp_class3/memory_network.py

@@ -425,7 +425,7 @@ def hop(query, story):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()

nlp_class3/poetry.py

@@ -160,8 +160,8 @@
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()
 

nlp_class3/wseq2seq.py

@@ -165,7 +165,8 @@
 # assign the values
 for i, d in enumerate(decoder_targets):
   for t, word in enumerate(d):
-    decoder_targets_one_hot[i, t, word] = 1
+    if word != 0:
+      decoder_targets_one_hot[i, t, word] = 1
 
 
 
@@ -190,7 +191,7 @@
 
 # this word embedding will not use pre-trained vectors
 # although you could
-decoder_embedding = Embedding(num_words_output, LATENT_DIM)
+decoder_embedding = Embedding(num_words_output, EMBEDDING_DIM)
 decoder_inputs_x = decoder_embedding(decoder_inputs_placeholder)
 
 # since the decoder is a "to-many" model we want to have
@@ -263,8 +264,8 @@ def acc(y_true, y_pred):
 plt.show()
 
 # accuracies
-plt.plot(r.history['acc'], label='acc')
-plt.plot(r.history['val_acc'], label='val_acc')
+plt.plot(r.history['accuracy'], label='acc')
+plt.plot(r.history['val_accuracy'], label='val_acc')
 plt.legend()
 plt.show()