I am stuck in my keras multilabel problem. I got the hint to work with a custom data generator to create small batches and avoid memory issues.
I work with a csv file with IDs, Filenames and their corresponding labels (21 in total), which looks like this:
Filename label1 label2 label3 label4 ... ID
abc1.jpg 1 0 0 1 ... id-1
def2.jpg 1 0 0 1 ... id-2
ghi3.jpg 1 0 0 1 ... id-3
...
I put the the ids and the labels in dictionaries which has the following output:
partition: {'train': ['id-1','id-2','id-3',...], 'validation': ['id-7','id-14','id-21',...]}
labels: {'id-0': [1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1],
'id-1': [1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1],
'id-2': [1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1],
...}
As well i have a folder with every image saved as a single npy file, which will be taken by the custom data generator below:
import numpy as np
import keras
from keras.layers import *
from keras.models import Sequential
class DataGenerator(keras.utils.Sequence):
'Generates data for Keras'
def __init__(self, list_IDs, labels, batch_size=32, dim=(224,224), n_channels=3,
n_classes=21, shuffle=True):
'Initialization'
self.dim = dim
self.batch_size = batch_size
self.labels = labels
self.list_IDs = list_IDs
self.n_channels = n_channels
self.n_classes = n_classes
self.shuffle = shuffle
self.on_epoch_end()
def __len__(self):
'Denotes the number of batches per epoch'
return int(np.floor(len(self.list_IDs) / self.batch_size))
def __getitem__(self, index):
'Generate one batch of data'
# Generate indexes of the batch
indexes = self.indexes[index*self.batch_size:(index+1)*self.batch_size]
# Find list of IDs
list_IDs_temp = [self.list_IDs[k] for k in indexes]
# Generate data
X, y = self.__data_generation(list_IDs_temp)
return X, y
def on_epoch_end(self):
'Updates indexes after each epoch'
self.indexes = np.arange(len(self.list_IDs))
if self.shuffle == True:
np.random.shuffle(self.indexes)
def __data_generation(self, list_IDs_temp):
'Generates data containing batch_size samples' # X : (n_samples, *dim, n_channels)
# Initialization
X = np.empty((self.batch_size, *self.dim, self.n_channels))
y = np.empty((self.batch_size), dtype=int)
# Generate data
for i, ID in enumerate(list_IDs_temp):
# Store sample
X[i,] = np.load('Folder with npy files/' + ID + '.npy')
# Store class
y[i] = self.labels[ID]
return X, keras.utils.to_categorical(y, num_classes=self.n_classes)
import numpy as np
from keras.models import Sequential
# Parameters
params = {'dim': (224, 224),
'batch_size': 32,
'n_classes': 21,
'n_channels': 3,
'shuffle': True}
# Datasets
partition = partition
labels = labels
# Generators
training_generator = DataGenerator(partition['train'], labels, **params)
validation_generator = DataGenerator(partition['validation'], labels, **params)
# Design model
model = Sequential()
model.add(Conv2D(32, (3,3), input_shape=(224, 224, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2,2)))
...
model.add(Flatten())
model.add(Dense(64))
model.add(Activation('relu'))
model.add(Dense(21))
model.add(Activation('softmax'))
model.summary()
so far my notebook doesnt give me any Errors, but when i do the following:
model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy'])
# Train model on dataset
model.fit_generator(generator=training_generator,
validation_data=validation_generator,
epochs=5,
use_multiprocessing=True,
workers=2)
i get an Error message like this:
Exception in thread Thread-7: Traceback (most recent call last): File "c:\users\sebas\appdata\local\programs\python\python36\lib\threading.py", line 916, in _bootstrap_inner self.run() ...
File "c:\users\sebas\appdata\local\programs\python\python36\lib\multiprocessing\reduction.py", line 60, in dump ForkingPickler(file, protocol).dump(obj) BrokenPipeError: [Errno 32] Broken pipe
It feels that I pass or use the data somehow incorrect!? If someone has an idea or a hint how to pass the data in a better manner or tackle this issue, i will appreciate. Even a different approach would be awesome. Thanks for your help in advance.
