TextCNN_Torch have wrong comment

graykode / nlp-tutorial

Natural Language Processing Tutorial for Deep Learning Researchers

MIT License

14.29k stars 3.95k forks source link

Open jnakor opened 4 years ago

jnakor commented 4 years ago

def forward(self, X): embedded_chars = self.W[X] # [batch_size, sequence_length, sequence_length]

I think the shape is [batch_size, sequence_length,embedding_size]

endeavor11 commented 4 years ago

yes，I think so

Yuhuishishishi commented 4 years ago

Filed PR #49

AgaigetS commented 4 years ago

can somebody tell me, why need three conv layer to convolve the word embedding matrix? I dont understand.