A wrong comment in tdnn about the dim of conv

mkroutikov / tf-lstm-char-cnn

tensorflow rewrite of Yoon Kim's lstm-char-cnn

MIT License

215 stars 74 forks source link

Closed yifeiwang77 closed 6 years ago

yifeiwang77 commented 6 years ago

the comment in line 95 says

# [batch_size x max_word_length x embed_size x kernel_feature_size]

, which I think actually should be

[batch_size x 1 x reduced_length x kernel_feature_size]

That confused me a lot as I'm a beginner to TF.

And you also seem to omit all the num_unroll_steps in the comments of this part.

mkroutikov commented 6 years ago

yes, comments are obsolete... some were leftovers from the parent project.

sorry if that caused you a confusion.

could you make a PR that fixes the comments?

mkroutikov commented 6 years ago

Fixed