-
Notifications
You must be signed in to change notification settings - Fork 10
Issues: Modalities/modalities
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Author
Label
Projects
Milestones
Assignee
Sort
Issues list
FSDP2: Full Activation Checkpointing does not work with torch.compile
bug
Something isn't working
#361
opened Apr 27, 2025 by
le1nux
Checkpoint conversion fails with the FSDP2 integration
bug
Something isn't working
#358
opened Apr 18, 2025 by
le1nux
Weight tying breaks with meta device initialisation
bug
Something isn't working
#357
opened Apr 18, 2025 by
le1nux
GPU Peak performance calculation requires FP8 and GH200 support
enhancement
New feature or request
#353
opened Apr 16, 2025 by
le1nux
Weight Tying raises errors when running distributed checkpointing
#348
opened Apr 13, 2025 by
le1nux
Pipeline parallelism requires special gradient clipping implementation
bug
Something isn't working
#313
opened Mar 3, 2025 by
le1nux
Unify handling of licenses of foreign code
documentation
Improvements or additions to documentation
#308
opened Feb 24, 2025 by
BlueCrescent
Refactor Binary File Operations
enhancement
New feature or request
#294
opened Jan 17, 2025 by
mali-git
Limit the number of samples during training/evaluation by configuration
enhancement
New feature or request
#279
opened Dec 5, 2024 by
le1nux
Redundant training config parameter ffn_hidden in the case of SwiGLU
#276
opened Dec 4, 2024 by
flxst
Increase model_max_length to avoid tokenizer warnings during packing
#275
opened Dec 4, 2024 by
flxst
Consistency issue with max_length flag in tokenizer
bug
Something isn't working
#274
opened Dec 4, 2024 by
le1nux
Previous Next
ProTip!
Exclude everything labeled
bug
with -label:bug.