recurrent_network.py

[kecaiwu]

The comments of x=tf.split(0,n_steps,x) is 'split to get a list of 'n_steps' tensors of shape(batch_size,n_hidden)', the shape is (batch_sizs,n_hidden)? why not the (batch_size,n_input). Do i missing something?

[aymericdamien]

Yes, it is a typo, it should be (batch_size,n_input), this is fixed now :)