Jingjing Gong

icon company Tsinghua University icon location Haidian district, Beijing, China icon location 龚经经

Results 2 issues of Jingjing Gong

weight normalization is probably implemented wrong

1 comment

https://github.com/tobyyouup/conv_seq2seq/blob/78a6e4e62a4c57a5caa9d584033a85e810fd726e/seq2seq/encoders/conv_encoder_utils.py#L59 you are not using `V_norm` any where except from initialization of variable tensor `g`. and according to the paper https://arxiv.org/abs/1602.07868, g should be a scalar instead of a variable...

Training loss goes down and then go up

9 comment

I am using part of your code, mainly `conv_encoder_stack` , to encode a sentence. then I found it weird that the training loss would go down at first then go...