lucidrains/h-transformer-1d issues and pull requests

#25 - error in test

Issue - State: closed - Opened by jizhang02 8 months ago - 1 comment

#24 - Billion Word Benchmark - Reproducibility

Issue - State: open - Opened by DavidHerel about 1 year ago

#23 - Use for relatively short sequences and small datasets

Issue - State: open - Opened by Vedasheersh over 1 year ago

#22 - Number of processed inputs in causal version is off

Issue - State: open - Opened by vladyorsh over 1 year ago - 2 comments

#21 - Fix indexing

Pull Request - State: open - Opened by jglaser over 2 years ago - 1 comment

#20 - Approximated values are off

Issue - State: open - Opened by jglaser over 2 years ago - 2 comments

#19 - add key_value_mask argument for cross attention

Pull Request - State: closed - Opened by jglaser almost 3 years ago - 5 comments

#18 - Masking not working in training, thanks

Issue - State: closed - Opened by junyongyou almost 3 years ago - 6 comments

#17 - Sequence classfication, thanks a lot

Issue - State: closed - Opened by junyongyou almost 3 years ago

#16 - Add Norm Missing

Issue - State: closed - Opened by wwx13 almost 3 years ago - 2 comments

#15 - Mask not working

Issue - State: closed - Opened by wwx13 about 3 years ago - 2 comments

#14 - Multiple applications of positional embeddings?

Issue - State: closed - Opened by Clemens123 about 3 years ago - 1 comment

#13 - Algorithm Mismatch

Issue - State: closed - Opened by jinmang2 about 3 years ago - 3 comments

#12 - Application to sequence classification?

Issue - State: closed - Opened by trpstra about 3 years ago - 4 comments

#11 - Does its model include relative position embedding?

Issue - State: closed - Opened by hadaev8 about 3 years ago - 1 comment

#10 - One simple question

Issue - State: closed - Opened by CiaoHe about 3 years ago - 2 comments

#9 - eos token does not work in batch mode generation

Issue - State: closed - Opened by tmphex about 3 years ago - 4 comments

#8 - Sequence length issue when `causal = True`

Issue - State: closed - Opened by jaak-s about 3 years ago - 1 comment

#7 - Does h-transformer-1d have a subquadratic cost?

Issue - State: closed - Opened by jinmang2 about 3 years ago - 1 comment

#6 - RuntimeError: Tensor type unknown to einops <class 'torch.return_types.max'>

Issue - State: closed - Opened by wajihullahbaig about 3 years ago - 4 comments

#5 - ModuleNotFoundError: No module named 'rotary_embedding_torch'

Issue - State: closed - Opened by wajihullahbaig about 3 years ago - 1 comment

#4 - Mini-batching (b > 1) does not work with masking

Issue - State: closed - Opened by jaak-s about 3 years ago - 2 comments

#3 - Example in README does not work