Skip to content

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.

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

100 Open 1,564 Closed
100 Open 1,564 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

Support int16?
#3295 opened Dec 15, 2024 by Bing-a-ling7
How to run accelerate with PYTORCH_ENABLE_MPS_FALLBACK
#3294 opened Dec 15, 2024 by mirodil-ml
4 tasks
Adam update occurs during accelerator.backward rather than optimizer.step
#3291 opened Dec 12, 2024 by jsmith-1989
RuntimeError: Input type (float) and bias type (c10::Half) should be the same
#3290 opened Dec 11, 2024 by PABannier
2 of 4 tasks
report: "ValueError: Batch does not contain any data (None)." when streaming=True
#3286 opened Dec 10, 2024 by holyYodu
2 of 4 tasks
Question About Gradient Synchronization During the Accelerate Process
#3285 opened Dec 10, 2024 by Klein-Lan
4
[FR] Support config torch.compile()/TorchDynamoPlugin with dynamic=None
#3284 opened Dec 9, 2024 by IrineSistiana
TypeError: Accelerator.__init__() got an unexpected keyword argument 'kwargs_handler'
#3283 opened Dec 9, 2024 by fanfanffff1
2 of 4 tasks
1
Accelerate.unwrap should fully unwrap the model.
#3281 opened Dec 9, 2024 by ggoggam
2 of 4 tasks
Some Questions About 'accelerate.prepare()'
#3278 opened Dec 7, 2024 by Klein-Lan
PaliGemma2 peft + accelerate evaluation during training fails with TypeError: Unsupported types (<class 'transformers.cache_utils.HybridCache'>) passed to _pad_across_processes
#3277 opened Dec 7, 2024 by beniz
1 of 4 tasks
1
Multi-Node (2x8 H100) training progress bar not showing up! (Node Connection Issue)
#3273 opened Dec 6, 2024 by KeshavSingh29
2 of 4 tasks
device_map=balanced not support in NPU when using FLUX.1-dev model
#3271 opened Dec 4, 2024 by baymax591
2 of 4 tasks
1
🤨Question: What if model has float16 dtype and mixed_precision is set to fp16 as well?
#3269 opened Nov 29, 2024 by townwish4git
[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
accelerator.prepare
#3265 opened Nov 26, 2024 by wfan1203
Multi-GPU DeepSpeed Stage-3 doesn't reduce memory compared to Single GPU
#3264 opened Nov 26, 2024 by enesmsahin
2 of 4 tasks
How to load checkpoint shards one by one to avoid OOM error?
#3263 opened Nov 26, 2024 by amoyplane
2 of 4 tasks
1
How to Properly Resume Multi-GPU Training with accelerate launch Without OOM or Loss Issues?
#3260 opened Nov 25, 2024 by tqxg2018
Non-root FSDP instance's _is_root should not have been set yet or should have been set to False while saving state with FSDP
#3258 opened Nov 25, 2024 by enesmsahin
2 of 4 tasks
# [BUG] [Fix-Suggested] Model Training Stalls with FSDP when fsdp_use_orig_params=False due to inconsistent model-optimizer state
#3256 opened Nov 24, 2024 by traincheck-team
2
ModuleNotFoundError: No module named 'torchvision'
#3254 opened Nov 23, 2024 by Zerycii
2 of 4 tasks
1
Accelerate + FSDP plugin hang on after model save intermediate checkpoint
#3250 opened Nov 22, 2024 by leeruibin
2 of 4 tasks
1
examples/inference/pippy/llama.py Assertion error about graphs
#3249 opened Nov 22, 2024 by 685Degrees
2 of 4 tasks
🚀 Feature Request: Improve stateful_dataloader by passing snapshot_every_n_steps
#3243 opened Nov 18, 2024 by yzhangcs
ProTip! Find all open issues with in progress development work with linked:pr.