Styjun

How should I write this passage to make it the most readable?

Why aren't rockets built with truss structures inside their fuel & oxidizer tanks to increase structural strength?

Is this n-speak?

Graphs for which a calculus student can reasonably compute the arclength

How do I call a 6-digit Australian phone number with a US-based mobile phone?

Why is there a large performance impact when looping over an array over 240 elements?

Luggage Storage at Szechenyi Baths

Creating some gif with tikz: Any idea to get better result?

Does fossil fuels use since 1990 account for half of all the fossil fuels used in history?

Are there any lower-level means of travelling between planes of existence?

Causal Diagrams using Wolfram?

Why aren't rainbows blurred-out into nothing after they are produced?

How far did Gandalf and the Balrog drop from the bridge in Moria?

Are there any cons in using rounded corners for bar graphs?

What is the safest way to leave my MacBook on 24/7 with macOS Mojave?

Escape Velocity - Won't the orbital path just become larger with higher initial velocity?

Why command hierarchy, if the chain of command is standing next to each other?

How much can I judge a company based on a phone screening?

A trip to the library

What kind of liquid can be seen 'leaking' from the upper surface of the wing of a Boeing 737-800?

How was the murder committed?

"Mouth-breathing" as slang for stupidity

Pokemon Go: Gym Badge Over-completed?

Help, I cannot decide when to start the story

Cannot add tensor to the batch: number of elements does not match

noisy validation loss (versus epoch) when using batch normalizationparameters value in tensorflowCombining 2D CNN with GRU in KerasObject center detection using Convnet is always returning center of image rather than center of objectTensorflow - ValueError: Rank mismatch:Implementing a generator for keras results in worse resultsEpoch's steps taking too long on GPUThe output NN is image an image with values 0 or 1, but the expected are a range of integers between 0 and 255Is it possible to train a CNN starting at an intermediate layer (in general and in Keras)?'Sequential' object has no attribute 'loss' - When I used GridSearchCV to tuning my Keras model

.everyoneloves__top-leaderboard:empty,.everyoneloves__mid-leaderboard:empty,.everyoneloves__bot-mid-leaderboard:empty margin-bottom:0;

1

I am experiencing a problem while training a neuronal network consisting of two convolution layers. I have a bunch of images that I normalize and resize to shape (28,28,3) within an "make iterator" function.

The training phase runs then normally till I encounter this error :

"InvalidArgumentError: Cannot add tensor to the batch: number of elements does not match. Shapes are: [tensor]: [28,28,4], [batch]: [28,28,3]
[[node IteratorGetNext_31]]"

I first tried to load some images and resize them using opencv and it then worked. However given the size of my dataset I need to use iterators and dataset objects provided by keras util to train my network on all the images I have.

Here is the code for the iterator:

def make_iterator(filenames, labels, batch_size):
 dataset = tf.data.Dataset.from_tensor_slices((filenames, labels))

 def parse(filename, label):
 image = tf.read_file(filename)
 image = tf.image.decode_jpeg(image)
 image = tf.cast(image, tf.float32)
 image = tf.image.resize(image, (28,28))
 image = tf.reshape(image, [28,28,3])
 image = image / 256
 return 'image': image, 'label': label

 dataset = dataset.apply(tf.data.experimental.map_and_batch(
 map_func=parse, batch_size=batch_size, num_parallel_batches=8))

 return dataset.make_one_shot_iterator()

Here are the details of my Convnet:

model = keras.Sequential()
model.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 3)))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(2, activation='softmax'))

And last here are the parameters for the fit phase:

history = model.fit(images
 , labels
 , epochs=200
 , steps_per_epoch= len(train_image_labels) // batch_size
 )

Thanks in advance :)

keras deep-learning iterator batch-processing

share|improve this question

asked Mar 27 at 10:52

mthanmthan

633 bronze badges

add a comment | 

1

I am experiencing a problem while training a neuronal network consisting of two convolution layers. I have a bunch of images that I normalize and resize to shape (28,28,3) within an "make iterator" function.

The training phase runs then normally till I encounter this error :

"InvalidArgumentError: Cannot add tensor to the batch: number of elements does not match. Shapes are: [tensor]: [28,28,4], [batch]: [28,28,3]
[[node IteratorGetNext_31]]"

I first tried to load some images and resize them using opencv and it then worked. However given the size of my dataset I need to use iterators and dataset objects provided by keras util to train my network on all the images I have.

Here is the code for the iterator:

def make_iterator(filenames, labels, batch_size):
 dataset = tf.data.Dataset.from_tensor_slices((filenames, labels))

 def parse(filename, label):
 image = tf.read_file(filename)
 image = tf.image.decode_jpeg(image)
 image = tf.cast(image, tf.float32)
 image = tf.image.resize(image, (28,28))
 image = tf.reshape(image, [28,28,3])
 image = image / 256
 return 'image': image, 'label': label

 dataset = dataset.apply(tf.data.experimental.map_and_batch(
 map_func=parse, batch_size=batch_size, num_parallel_batches=8))

 return dataset.make_one_shot_iterator()

Here are the details of my Convnet:

model = keras.Sequential()
model.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 3)))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(2, activation='softmax'))

And last here are the parameters for the fit phase:

history = model.fit(images
 , labels
 , epochs=200
 , steps_per_epoch= len(train_image_labels) // batch_size
 )

Thanks in advance :)

keras deep-learning iterator batch-processing

share|improve this question

asked Mar 27 at 10:52

mthanmthan

633 bronze badges

add a comment | 

I am experiencing a problem while training a neuronal network consisting of two convolution layers. I have a bunch of images that I normalize and resize to shape (28,28,3) within an "make iterator" function.

The training phase runs then normally till I encounter this error :

"InvalidArgumentError: Cannot add tensor to the batch: number of elements does not match. Shapes are: [tensor]: [28,28,4], [batch]: [28,28,3]
[[node IteratorGetNext_31]]"

I first tried to load some images and resize them using opencv and it then worked. However given the size of my dataset I need to use iterators and dataset objects provided by keras util to train my network on all the images I have.

Here is the code for the iterator:

def make_iterator(filenames, labels, batch_size):
 dataset = tf.data.Dataset.from_tensor_slices((filenames, labels))

 def parse(filename, label):
 image = tf.read_file(filename)
 image = tf.image.decode_jpeg(image)
 image = tf.cast(image, tf.float32)
 image = tf.image.resize(image, (28,28))
 image = tf.reshape(image, [28,28,3])
 image = image / 256
 return 'image': image, 'label': label

 dataset = dataset.apply(tf.data.experimental.map_and_batch(
 map_func=parse, batch_size=batch_size, num_parallel_batches=8))

 return dataset.make_one_shot_iterator()

Here are the details of my Convnet:

model = keras.Sequential()
model.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 3)))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(2, activation='softmax'))

And last here are the parameters for the fit phase:

history = model.fit(images
 , labels
 , epochs=200
 , steps_per_epoch= len(train_image_labels) // batch_size
 )

Thanks in advance :)

keras deep-learning iterator batch-processing

share|improve this question

asked Mar 27 at 10:52

mthanmthan

633 bronze badges

def make_iterator(filenames, labels, batch_size):
 dataset = tf.data.Dataset.from_tensor_slices((filenames, labels))

 def parse(filename, label):
 image = tf.read_file(filename)
 image = tf.image.decode_jpeg(image)
 image = tf.cast(image, tf.float32)
 image = tf.image.resize(image, (28,28))
 image = tf.reshape(image, [28,28,3])
 image = image / 256
 return 'image': image, 'label': label

 dataset = dataset.apply(tf.data.experimental.map_and_batch(
 map_func=parse, batch_size=batch_size, num_parallel_batches=8))

 return dataset.make_one_shot_iterator()

model = keras.Sequential()
model.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 3)))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(2, activation='softmax'))

history = model.fit(images
 , labels
 , epochs=200
 , steps_per_epoch= len(train_image_labels) // batch_size
 )

share|improve this question

asked Mar 27 at 10:52

mthanmthan

633 bronze badges

share|improve this question

asked Mar 27 at 10:52

mthanmthan

633 bronze badges

asked Mar 27 at 10:52

mthanmthan

633 bronze badges

asked Mar 27 at 10:52

mthanmthan

633 bronze badges