ValueError: Input 0 is incompatible with layer conv_1: expected ndim=3, found ndim=4
Solution 1
In the documentation it is mentioned that we need to mention the input in a specific format which is (None,NumberOfFeatureVectors). In your case it will be (None,4)
https://keras.io/layers/convolutional/
When using this layer as the first layer in a model, provide an input_shape argument (tuple of integers or None, e.g. (10, 128) for sequences of 10 vectors of 128-dimensional vectors, or (None, 128) for variable-length sequences of 128-dimensional vectors.
Solution 2
Try to put the input to the network like this:
Input(shape=(None, 4)
Generally, it's for the case you don't know the length of your sequence, but I had the same problem and for some reason, it was solved when I did it
Hope it works!
Solution 3
Specify the kernel_size in your convolutional layers as a tuple, not integer, even if it requires only a dimension:
e = Conv1D(9, (9), activation = 'relu', name='conv_1')(seq)
Although in Keras documentation is stated that both an integer and a tuple are valid, I found the second to be more handy with dimensionality.
Benjamin Lee
My name is Benjamin D. Lee. I'm currently a senior at Harvard studying computer science with a secondary in organismic and evolutionary biology. At heart, I'm a software engineer with a passion for research. My primary interests are in computational biology, robotic chemistry, evolutionary computing, and alignment-free sequence analysis. I've done my research at In-Q-Tel's Lab41 for the last two years, although I also spent a year in Harvard Medical School's Church lab. For more information about the work I've done (all of which is open source), take a look at my projects and published papers.
Updated on March 25, 2020Comments
-
Benjamin Lee about 4 years
I am trying to make a variational auto encoder to learn to encode DNA sequences, but am getting an unexpected error.
My data is an array of one-hot arrays.
The issue I'm getting is a Value Error. It's telling me that I have a four dimensional input, when my input is clearly three-dimensional (100, 4008, 4).
In fact, when I print out the
seq
layer, it says that it's shape is (?, 100, 4008, 4).When I take out a dimension, it then gives me an error for being two dimensional.
Any help will be highly appreciated!
Code is:
from keras.layers import Input from keras.layers.convolutional import Conv1D from keras.layers.core import Dense, Activation, Flatten, RepeatVector, Lambda from keras import backend as K from keras.layers.wrappers import TimeDistributed from keras.layers.recurrent import GRU from keras.models import Model from keras import objectives from one_hot import dna_sequence_to_one_hot from random import shuffle import numpy as np # take FASTA file and convert into array of vectors seqs = [line.rstrip() for line in open("/home/ubuntu/sequences.fa", "r").readlines() if line[0] != ">"] seqs = [dna_sequence_to_one_hot(s) for s in seqs] seqs = np.array(seqs) # first random thousand are training, next thousand are validation test_data = seqs[:1000] validation_data = seqs[1000:2000] latent_rep_size = 292 batch_size = 100 epsilon_std = 0.01 max_length = len(seqs[0]) charset_length = 4 epochs = 100 def sampling(args): z_mean_, z_log_var_ = args # batch_size = K.shape(z_mean_)[0] epsilon = K.random_normal_variable((batch_size, latent_rep_size), 0., epsilon_std) return z_mean_ + K.exp(z_log_var_ / 2) * epsilon # loss function def vae_loss(x, x_decoded_mean): x = K.flatten(x) x_decoded_mean = K.flatten(x_decoded_mean) xent_loss = max_length * objectives.categorical_crossentropy(x, x_decoded_mean) kl_loss = - 0.5 * K.mean(1 + z_log_var - K.square(z_mean) - K.exp(z_log_var), axis = -1) return xent_loss + kl_loss # Encoder seq = Input(shape=(100, 4008, 4), name='one_hot_sequence') e = Conv1D(9, 9, activation = 'relu', name='conv_1')(seq) e = Conv1D(9, 9, activation = 'relu', name='conv_2')(e) e = Conv1D(9, 9, activation = 'relu', name='conv_3')(e) e = Conv1D(10, 11, activation = 'relu', name='conv_4')(e) e = Flatten(name='flatten_1')(e) e = Dense(435, activation = 'relu', name='dense_1')(e) z_mean = Dense(latent_rep_size, name='z_mean', activation = 'linear')(e) z_log_var = Dense(latent_rep_size, name='z_log_var', activation = 'linear')(e) z = Lambda(sampling, output_shape=(latent_rep_size,), name='lambda')([z_mean, z_log_var]) encoder = Model(seq, z) # Decoder d = Dense(latent_rep_size, name='latent_input', activation = 'relu')(z) d = RepeatVector(max_length, name='repeat_vector')(d) d = GRU(501, return_sequences = True, name='gru_1')(d) d = GRU(501, return_sequences = True, name='gru_2')(d) d = GRU(501, return_sequences = True, name='gru_3')(d) d = TimeDistributed(Dense(charset_length, activation='softmax'), name='decoded_mean')(d) # create the model, compile it, and fit it vae = Model(seq, d) vae.compile(optimizer='Adam', loss=vae_loss, metrics=['accuracy']) vae.fit(x=test_data, y=test_data, epochs=epochs, batch_size=batch_size, validation_data=validation_data)