尝试以频谱图的形式对音频进行分类时，收到错误消息“ValueError：没有为任何变量提供渐变：...”

Question

Im trying to classify spectograms of shape 40, 501 and there are 15 classes.

import numpy as np
import tensorflow as tf
from keras.utils import to_categorical
from keras.models import Sequential
from keras.layers import Conv2D, MaxPooling2D, Dense, Flatten


X_test = np.load("X_test.npy")
y_test = np.load("y_test.npy")
X_train = np.load("X_train.npy")
y_train = np.load("y_train.npy")
X_train = X_train[..., np.newaxis]
y_train = to_categorical(y_train, 15)

model = tf.keras.models.Sequential()
model.add(Conv2D(32, kernel_size=(3, 3), input_shape=(40, 501, 1), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(32, kernel_size=(3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Flatten())
model.add(Dense(15, activation='softmax'))

model.summary()
model.compile()
model.fit(X_train, y_train, epochs=5)

Training data consists of 4500 spectograms (the data is preprocessed and I havent touched it. When i try the model.fit Im getting an error message:

Error:

ValueError: No gradients provided for any variable: ['conv2d/kernel:0', 'conv2d/bias:0', 'conv2d_1/kernel:0', 'conv2d_1/bias:0', 'dense/kernel:0', 'dense/bias:0'].

Answer 1

This an issue with how you coded the data generation, probably not in Tensorflow. It should return a tuple or a dict of tuples in dataloader.

To resolve this you must modify the data generator function as:

yield([array(Ximages),array(XSeq)],array()) 

instead of:

yield [[array(Ximages), array(XSeq)], array(y)]

Sample structure:

# data generator, intended to be used in a call to model.fit_generator()
def data_generator(descriptions, features, n_step, max_sequence):
    # loop until we finish training
    while 1:
        # loop over photo identifiers in the dataset
        for i in range(0, len(descriptions), n_step):
            Ximages, XSeq, y = list(), list(),list()
            for j in range(i, min(len(descriptions), i+n_step)):
                image = features[j]
                # retrieve text input
                desc = descriptions[j]
                # generate input-output pairs
                in_img, in_seq, out_word = preprocess_data([desc], [image], max_sequence)
                for k in range(len(in_img)):
                    Ximages.append(in_img[k])
                    XSeq.append(in_seq[k])
                    y.append(out_word[k])

       ### —-> CHANGE THIS TO FIT YOUR DATA: yield this batch of samples to the model
            #yield [[array(Ximages), array(XSeq)], array(y)]
yield([array(Ximages),array(XSeq)],array()) 

#Create the network 
image_model = Sequential()
image_model.add(Conv2D(16, (3, 3), padding='valid', activation='relu', input_shape=(256, 256, 3,)))
image_model.add(Conv2D(16, (3,3), activation='relu', padding='same', strides=2))
image_model.add(Conv2D(32, (3,3), activation='relu', padding='same'))
image_model.add(Conv2D(32, (3,3), activation='relu', padding='same', strides=2))
image_model.add(Conv2D(64, (3,3), activation='relu', padding='same'))
image_model.add(Conv2D(64, (3,3), activation='relu', padding='same', strides=2))
image_model.add(Conv2D(128, (3,3), activation='relu', padding='same'))

...
### —> data generator is called on model fit 
model.fit_generator(data_generator(texts, train_features, 1, max_sequence), steps_per_epoch=steps, epochs=1, callbacks=callbacks_list, verbose=1)
model.save(mydrive + '/output/weights.hdf5')