Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / lucidrains/sinkhorn-transformer issues and pull requests
#29 - A wrapper of SinkhornTransformerEncDec
Issue -
State: open - Opened by halexan about 3 years ago
#28 - Error with Autopadder
Issue -
State: closed - Opened by soonchangAI over 3 years ago
- 2 comments
#27 - A noobish question about training...
Issue -
State: closed - Opened by asigalov61 almost 4 years ago
- 9 comments
#26 - No module named 'local_attention'
Issue -
State: closed - Opened by asigalov61 almost 4 years ago
- 2 comments
#25 - Some questions about dropout
Issue -
State: closed - Opened by tatp22 over 4 years ago
- 6 comments
#24 - Fixed bug with dim_head and dim
Pull Request -
State: closed - Opened by tatp22 over 4 years ago
- 1 comment
#23 - Positional Embedding
Issue -
State: closed - Opened by ilya16 over 4 years ago
- 4 comments
#22 - several question about implementation
Issue -
State: closed - Opened by w4-magnes over 4 years ago
- 9 comments
#21 - Training crashes due to inplace operation
Issue -
State: closed - Opened by NaxAlpha over 4 years ago
- 6 comments
#20 - Size mismatch error when loading saved weights
Issue -
State: closed - Opened by epetros over 4 years ago
- 2 comments
#19 - Sortcut variant and bucket size
Issue -
State: closed - Opened by jsmith1915 over 4 years ago
- 16 comments
#18 - Performance of the different attention variants in the repo
Issue -
State: closed - Opened by jsmith1915 over 4 years ago
- 7 comments
#17 - TypeError exception in AxialPositionalEncoding when using DataParallel
Issue -
State: closed - Opened by kl0211 almost 5 years ago
- 8 comments
#16 - add feature to be able to sort multiple buckets, using multiple sorti…
Pull Request -
State: closed - Opened by lucidrains almost 5 years ago
- 1 comment
#15 - generation problem in a toy task
Issue -
State: closed - Opened by py4 almost 5 years ago
- 18 comments
#14 - dimension mismatch in context-only attention when seq_in and seq_out have different lengths
Issue -
State: closed - Opened by py4 almost 5 years ago
- 3 comments
#13 - axial positional embedding
Issue -
State: closed - Opened by rouniuyizu almost 5 years ago
- 2 comments
#12 - AttributeError: 'ProjectInOut' object has no attribute 'pad_to_bucket_size'
Issue -
State: closed - Opened by timsoraro almost 5 years ago
- 4 comments
#11 - AssertionError: different bucketing for key/values for causal reordering not supported yet
Issue -
State: closed - Opened by timsoraro almost 5 years ago
- 4 comments
#10 - make flexible sequence length work, and make model compatible with Au…
Pull Request -
State: closed - Opened by lucidrains almost 5 years ago
#9 - RuntimeError: range cannot be used as a value
Issue -
State: closed - Opened by timsoraro almost 5 years ago
- 2 comments
#8 - Do I need to pad?
Issue -
State: closed - Opened by timsoraro almost 5 years ago
- 8 comments
#7 - Update sinkhorn_transformer.py
Pull Request -
State: closed - Opened by blizda almost 5 years ago
#6 - complete input and context masking
Pull Request -
State: closed - Opened by lucidrains almost 5 years ago
#5 - add local attention
Pull Request -
State: closed - Opened by lucidrains almost 5 years ago
#4 - Training falling on version 0.0.14 and 0.0.15
Issue -
State: closed - Opened by blizda almost 5 years ago
- 30 comments
#3 - Post pre-trained models please
Issue -
State: closed - Opened by Hellisotherpeople almost 5 years ago
- 1 comment
#2 - RuntimeError: range of [nan, nan] is not finite
Issue -
State: closed - Opened by blizda almost 5 years ago
- 3 comments
#1 - Sequence Masking
Issue -
State: closed - Opened by calclavia almost 5 years ago
- 3 comments