-
Notifications
You must be signed in to change notification settings - Fork 1.1k
Issues: huggingface/accelerate
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
[suggestion] potentially add warning with "invalid"
device_map keys
#3584
opened May 22, 2025 by
johannaSommer
Multi gpu setup got stuck with accelerate but torchrun works
#3568
opened May 13, 2025 by
Jason3900
2 of 4 tasks
There might be a bug when using DeepSpeed with gradient_accumulation_steps > 1
#3552
opened May 9, 2025 by
hiawui
This can potentially cause a hang if this rank to GPU mapping is incorrect.
#3551
opened May 9, 2025 by
Unturned3
3 of 4 tasks
datasets Iterable Dataset sharding support
feature request
#3547
opened May 5, 2025 by
ValMystletainn
RuntimeError: Tensor on device meta is not on the expected device cuda:0!
good first issue
Good for newcomers
#3542
opened Apr 30, 2025 by
nitinmukesh
2 of 4 tasks
Error loading optimizer state with torch.load weights_only=True default in PyTorch 2.6
#3539
opened Apr 29, 2025 by
luiz0992
2 of 4 tasks
NCCL timeout when try to save
bug
Something isn't working
good first issue
Good for newcomers
#3512
opened Apr 18, 2025 by
thechaos16
2 of 4 tasks
FSDP2 Unwrap model, still has dtensors as weights
bug
Something isn't working
#3487
opened Apr 5, 2025 by
kiansierra
2 of 4 tasks
DeepSpeed ZeRO Stage 2/3 incompatibility with no_sync context manager
#3481
opened Apr 4, 2025 by
alisafaya
[Feature Request] add a memory logging manager
contributions-welcome
feature request
Request for a new feature to be added to Accelerate
#3300
opened Dec 17, 2024 by
sayakpaul
[CPU offload hooks] hooks with overlapped transfers and computations
feature request
Request for a new feature to be added to Accelerate
#3267
opened Nov 29, 2024 by
sayakpaul
[BUG] Accelerate 1.0.1 failed to train multiple zero-3 models
wip
Work in progress
#3170
opened Oct 15, 2024 by
nrailg
2 of 4 tasks
Impossibility to use num_workers and prefetch_factor when using StatefulDataLoader (use_stateful_dataloader=True)
enhancement
New feature or request
feature request
Request for a new feature to be added to Accelerate
#3110
opened Sep 13, 2024 by
hkproj
2 of 4 tasks
[Community Contributions] examples on distributed inference using 🤗 Accelerate
contributions-welcome
good first issue
Good for newcomers
wip
Work in progress
#3078
opened Sep 4, 2024 by
sayakpaul
infer_auto_device_map inefficiently allocates GPU memory for models with imbalanced module sizes
wip
Work in progress
#3041
opened Aug 25, 2024 by
Nech-C
2 of 4 tasks
Barebones dataloader to allow for any type of iterable dataloader-like object to be used. Should just handle device placement
good first issue
Good for newcomers
#2975
opened Jul 29, 2024 by
muellerzr
Previous Next
ProTip!
Mix and match filters to narrow down what you’re looking for.