Transformer_Relative_Position_PyTorch
Transformer_Relative_Position_PyTorch copied to clipboard
Implement the paper "Self-Attention with Relative Position Representations"
example hypothesis: this man was mahhhhnenearararari. example hypothesis: if we can stand in the next 150 years, i think our ururururururge. I train 30epoch but only get 11BLUE.... sould I...
Hi, Thanks a lot for putting this implementation out there. It helped me a lot! I was just curious why the `contiguous()` was called after `permute` or `transpose` but always...
显存溢出
使用Relative Position后显存占用明显增加是什么原因呢?
Hi @evelinehong , could you please kindly add a license to this repo, e.g., MIT license? Thanks a lot!