Skip to content

Issues: facebookresearch/fairscale

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.

Sign up for GitHub

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

Clear current search query, filters, and sorts
104 Open 1,086 Closed
104 Open 1,086 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

AttributeError: 'FlatParameter' object has no attribute '_full_param_padded'
#1193 opened Jan 19, 2025 by hui-li-xf
When mixed precision input contains grad tensor, FSDP cast it with no grad
#1191 opened Dec 8, 2024 by kamwoh
support for grad acc CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1190 opened Aug 30, 2024 by ngoyal2707 Loading…
1
Hi, Groups division may be incorrect in initialize() in fairscale/nn/model_parallel/initialize.py
#1189 opened Aug 30, 2024 by Youngluc
Raising assert param.grad is not None when finetuning LoRA.
#1188 opened Jun 26, 2024 by HashimotoPatrickMu
1
Bump scikit-learn from 1.1.3 to 1.5.0 CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. dependencies Pull requests that update a dependency file
#1187 opened Jun 18, 2024 by dependabot bot Loading…
FP8 AllGather Support in Fairscale CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1185 opened May 20, 2024 by levendlee Loading…
10 tasks
6
Llama4 FP8 Training Debug - fairscale CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1183 opened May 7, 2024 by jiecaoyu Loading…
10 tasks
Fix minor grammatical corrections in docs CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1181 opened Apr 29, 2024 by aakashapoorv Loading…
6 of 10 tasks
[FSDPv1] Only perform cat() during last microbatch backward() within FlattenParamsWrapper CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1180 opened Apr 29, 2024 by chrisxcai • Draft
4
[WIP] Make FSDPv1 only perform cat() during last microbatch backward() within FlattenParamsWrapper CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1178 opened Apr 19, 2024 by chrisxcai • Draft
1
Make sure that tensor is contiguous before gathering across processes CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1173 opened Apr 1, 2024 by patrickvonplaten Loading…
[question] Different training between DDP & Sharded DDP
#1172 opened Mar 29, 2024 by kwohlfahrt
Bump black from 22.3.0 to 24.3.0 CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. dependencies Pull requests that update a dependency file
#1169 opened Mar 20, 2024 by dependabot bot Loading…
Fairscale support for only performing allreduce in last microbatch CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1168 opened Mar 18, 2024 by jiecaoyu • Draft
10 tasks
3
Fix params_with_grad in FSDP when the model has frozen parameters CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1167 opened Mar 4, 2024 by whbldhwj Loading…
1 of 10 tasks
Example of MOE
#1165 opened Feb 28, 2024 by Juanhui28
1
Avoid calling _free_fp16_param_shard() too early CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1164 opened Feb 20, 2024 by jiecaoyu Loading…
10 tasks
3
Added reshard hook for frozen params in backward CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1159 opened Jan 12, 2024 by awgu Loading…
1 of 10 tasks
6
Add support for torch.set_default_device when initializing model parameters CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1158 opened Dec 28, 2023 by fshp971 Loading…
6 of 10 tasks
How can I use torchrun + model parallelism + FSDP
#1155 opened Dec 8, 2023 by HackGiter
1
Fix _free_full_params() CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1149 opened Nov 21, 2023 by hadasah Loading…
10 tasks
Extend CheckpointFunction to track all tensor input/output CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1148 opened Nov 7, 2023 by 000Justin000 Loading…
6 of 10 tasks
[Not for merge] fp8allgather debug CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#1147 opened Oct 29, 2023 by jiecaoyu Loading…
It is dangerous to using default non_block=True.
#1146 opened Oct 27, 2023 by heshenghuan
ProTip! Exclude everything labeled bug with -label:bug.