Skip to content

Pull requests: microsoft/DeepSpeed

New pull request New
129 Open 2,961 Closed
129 Open 2,961 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

Add arctic model support by adding w2 to all_reduce
#6856 opened Dec 11, 2024 by pi314ever Loading…
1
Fix assertion for offloading states
#6855 opened Dec 11, 2024 by tohtana Loading…
Fix: ensure deepspeed.initialize can only be called once.
#6848 opened Dec 10, 2024 by traincheck-team Loading…
1
2
Use ds-specific module id to avoid conflicts
#6847 opened Dec 10, 2024 by tjruwase Loading…
1
[inf] Add config var to enable keeping module on host
#6846 opened Dec 10, 2024 by oelayan7 Loading…
1
Update real_accelerator.py
#6845 opened Dec 10, 2024 by keiwoo Loading…
8
Cleanup/fix errors in ops/transformer/inference tests
#6830 opened Dec 6, 2024 by loadams • Draft
Add MLP/lm_head tp grain size setting.
#6828 opened Dec 6, 2024 by Yejing-Lai Loading…
5
Remove pin from transformers version and fix Processing/Threading issues in tests
#6822 opened Dec 5, 2024 by loadams Loading…
Support pure meta model lm_head tp
#6812 opened Dec 2, 2024 by Yejing-Lai Loading…
Avoid poisoning process with CUDA calls as soon as importing
#6810 opened Nov 29, 2024 by HollowMan6 Loading…
6
Zero2: avoid graph breaks in torch.compile by using param_idx
#6803 opened Nov 28, 2024 by nelyahu Loading…
3
Remove warnings from autodoc and sphinx
#6788 opened Nov 26, 2024 by loadams Loading…
Stage3: Use new torch grad accumulation hooks API
#6773 opened Nov 21, 2024 by deepcharm Loading…
4
Check transformers version in BLOOM for inference v1
#6766 opened Nov 19, 2024 by lekurile Loading…
1
BLOOM fixes for DS Legacy Inference
#6765 opened Nov 19, 2024 by lekurile • Draft
Add the missing view operations from sequence parallel(async).
#6750 opened Nov 14, 2024 by inkcherry Loading…
3
Fix building on Windows with presence of Triton
#6749 opened Nov 14, 2024 by woct0rdho Loading…
2
Update flake8 version
#6740 opened Nov 11, 2024 by loadams Loading…
Update formatting workflow
#6738 opened Nov 11, 2024 by loadams Loading…
Training ops kernels: Speeding up the Llama-based MoE architectures
#6734 opened Nov 8, 2024 by RezaYazdaniAminabadi • Draft
1
Support latest transformers with DSChat
#6711 opened Nov 4, 2024 by loadams Loading…
Allow launcher to include --include=node3, not just --include=node3:1,2,3,4,5,6,7,8
#6698 opened Nov 1, 2024 by stephen-nju Loading…
1
9
Reduce the device bubble introduced by heavy loop synchronization in coalesced fetch/release(z3_leaf_module)
#6694 opened Oct 31, 2024 by inkcherry Loading…
Update MII tests to support transformers latest
#6686 opened Oct 29, 2024 by loadams Loading…
ProTip! Type g i on any issue or pull request to go back to the issue listing page.