Skip to content

Pull requests: huggingface/accelerate

New pull request New
27 Open 1,655 Closed
27 Open 1,655 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
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Use torch.distributed.checkpoint.state_dict.set_model_state_dict in load_checkpoint_in_model
#3432 opened Mar 10, 2025 by ringohoffman Loading…
2 of 5 tasks
fix "clear_device_cache" in hooks.py causing a CUDA runtime error in multiple GPUs environment.
#3430 opened Mar 9, 2025 by CyberVy Loading…
1
[Feat] Layerwise casting hook
#3427 opened Mar 7, 2025 by sayakpaul • Draft
3 tasks
1
Add log_artifact, log_artifacts and log_figure capabilities to the MLflowTracker.
#3419 opened Mar 3, 2025 by luiz0992 Loading…
5 tasks done
1
13
Distributed inference example for llava_next
#3417 opened Mar 1, 2025 by VladOS95-cyber Loading…
3 of 5 tasks
Fix AMD GPU support with should_reduce_batch_size()
#3405 opened Feb 21, 2025 by cameronshinn Loading…
5 tasks
1
Fix device KeyError in tied_params_map
#3403 opened Feb 20, 2025 by dvrogozh Loading…
1
1
xpu: enable xccl distributed backend
#3401 opened Feb 18, 2025 by dvrogozh Loading…
Support TP + FSDPv2 / HSDP or just FSDPv2 / HSDP
#3395 opened Feb 13, 2025 by kmehant • Draft
5 tasks
WIP: Initial FSDP2 support
#3394 opened Feb 11, 2025 by S1ro1 • Draft
5 tasks
11
tensor parallel dataloder for deepspeed accelerator
#3390 opened Feb 10, 2025 by inkcherry Loading…
10
allow to set mxied_precision in accelerate config when using a given deepspeed config
#3385 opened Feb 7, 2025 by XiaobingSuper Loading…
5 tasks
1
4
HPU support
#3378 opened Feb 4, 2025 by IlyasMoutawwakil Loading…
5 tasks
74
Fix command args for MPI based deepspeed launchers
#3357 opened Jan 20, 2025 by chiragjn Loading…
1 of 5 tasks
1
4
feat: Add no_ssh and slurm multinode launcher options for deepspeed
#3329 opened Jan 8, 2025 by hsmallbone Loading…
5 tasks done
8
Update CometMLTracker to allow re-using experiment
#3328 opened Jan 7, 2025 by Lothiraldan Loading…
1 of 5 tasks
9
[WIP] optimize infer_auto_device_map for multi-GPU allocation
#3321 opened Jan 2, 2025 by Nech-C • Draft
3 of 5 tasks
4
[FSDP] Make env var and dataclass flag consistent for cpu_ram_efficient_loading
#3307 opened Dec 19, 2024 by SumanthRH Loading…
1 of 5 tasks
5
[FSDP][Example] Fix FSDP example for use_orig_params=False
#3298 opened Dec 16, 2024 by SumanthRH Loading…
2 of 5 tasks
7
[RFC] Support FSDP2
#3231 opened Nov 8, 2024 by kmehant Loading…
1 of 5 tasks
1
11
add support for custom function for reducing the batch size wip Work in progress
#3071 opened Sep 3, 2024 by winglian Loading…
5 tasks
8
Some adjustment for supporting Deepspeed-Ulysses wip Work in progress
#2877 opened Jun 20, 2024 by zeyugao Loading…
1 of 5 tasks
17
handle weight sharing with init_on_device wip Work in progress
#2737 opened May 3, 2024 by aws-rhsoln Loading…
5 tasks
6
[DO NOT MERGE] add all level buffer support when computing infer_auto_device_map enhancement New feature or request feature request Request for a new feature to be added to Accelerate wip Work in progress
#2663 opened Apr 12, 2024 by SunMarc Loading…
3
make big model inference compatible with torch.compile wip Work in progress
#2609 opened Apr 2, 2024 by SunMarc • Draft
1
2
ProTip! Mix and match filters to narrow down what you’re looking for.