Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / hkproj/pytorch-transformer issues and pull requests
#31 - Fix Layernorm Implementation
Pull Request -
State: open - Opened by Xue10 about 2 months ago
#30 - Should encoder mask be a (1, seq_len, seq_len) matrix?
Issue -
State: open - Opened by scotthuang1989 about 2 months ago
#29 - Regarding add and norm block
Issue -
State: open - Opened by DevasishY 4 months ago
#28 - “Bus Error and Resource Tracker Warning When Training PyTorch Model on GPU with MPS”
Issue -
State: open - Opened by pratheeshkumar99 4 months ago
#27 - Inconsistency in model.decode() and forward method of Decode class
Issue -
State: open - Opened by SwastikGorai 5 months ago
#26 - Change
Pull Request -
State: open - Opened by NhatDinh2411 6 months ago
#25 - Error: DatasetGenerationError: An error occurred while generating the dataset
Issue -
State: open - Opened by hissain 6 months ago
#24 - Refactor Causal Mask Generation for Simplicity
Pull Request -
State: open - Opened by sanzgadea 7 months ago
#23 - W_o matrix in multiheadattention
Issue -
State: closed - Opened by PING-An32 7 months ago
#22 - Clarification regarding decoder_input and label
Issue -
State: closed - Opened by BrLlan 8 months ago
- 1 comment
#21 - Clarification regarding dropout in the multihead attention block
Issue -
State: open - Opened by anupsingh15 8 months ago
#20 - Not an issue, but a re-implementation
Issue -
State: open - Opened by chettiargautam 9 months ago
#19 - translate.py not consistent
Issue -
State: open - Opened by dzjxzyd 9 months ago
- 2 comments
#18 - only spaces get predicted
Issue -
State: closed - Opened by aadityapattabhiraman 9 months ago
- 3 comments
#17 - System cannot find the path specified - tokenizer path.
Issue -
State: closed - Opened by shuklaji28 10 months ago
- 1 comment
#16 - Getting Error : module transformers has no attribute 'PreTrainedtokenizerBase'
Issue -
State: closed - Opened by shuklaji28 10 months ago
#15 - Quality after 20 epoch training
Issue -
State: open - Opened by thanhnew2001 10 months ago
- 2 comments
#14 - The implementation of ResidualConnection is not correct?
Issue -
State: closed - Opened by imxtx 12 months ago
- 2 comments
#13 - Fourier Positional Embeddings
Issue -
State: open - Opened by mourad-ghafiri 12 months ago
#12 - Truncation for Tokenizer
Issue -
State: closed - Opened by ardaaras99 about 1 year ago
- 1 comment
#10 - forward method in transformer class
Issue -
State: closed - Opened by akkasi about 1 year ago
- 1 comment
#9 - dataset
Issue -
State: closed - Opened by debby-2020 about 1 year ago
- 3 comments
#8 - Helpful updates
Pull Request -
State: closed - Opened by ivolazy about 1 year ago
- 1 comment
#7 - projection_layer should output logits value directly instead of log_softmax
Issue -
State: closed - Opened by liyang85105 about 1 year ago
- 3 comments
#6 - Run encoder done.
Pull Request -
State: closed - Opened by jzh18 about 1 year ago
#5 - Question: train.py
Issue -
State: closed - Opened by mohamedelbahnasawi about 1 year ago
- 4 comments
#4 - multihead attention biases
Issue -
State: closed - Opened by stanchiang over 1 year ago
- 1 comment
#3 - clearing cache before epoch
Pull Request -
State: closed - Opened by kavinask007 over 1 year ago
- 1 comment
#2 - Position Encoding
Issue -
State: closed - Opened by Sheiphan over 1 year ago
- 2 comments
#1 - New dataset
Pull Request -
State: closed - Opened by hkproj over 1 year ago