Cao_enjun

return attn_dist / (tf.reshape(masked_sums, [-1, 1]) + tf.ones_like(tf.reshape(masked_sums, [-1, 1])) * sys.float_info.epsilon) tf.reshape(masked_sums, [-1, 1]) seems will not be zeros during training, but it is also harmful due to "division...

I change the model with the method in WGAN. I remove sigmoid function of the final layer in discriminator, and use RMSprop optimizer.

oh, I know ,it is same.

torch.cat([dict[word].view(1, -1) for word in lsent]) is this also ok ?

@lanwuwei Thanks for your help, I find my pytorch version is not 0.4.0.

Hi, I come back again. A question of virtual adversarial training arose to me. When using this method, we first feed the unlabeled data into the forward net to get...

I have the same issue, has someone solved？