Skip to content

Pull requests: pytorch/torchtitan

New pull request New
138 Open 1,180 Closed
138 Open 1,180 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
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Provide load_seed_checkpoint_only option CLA Signed This label is managed by the Meta Open Source bot.
#1800 opened Oct 5, 2025 by fegin Loading…
Add a loss comparison script CLA Signed This label is managed by the Meta Open Source bot.
#1799 opened Oct 5, 2025 by fegin Loading…
[RFC] Lift freqs_cis as an input of models CLA Signed This label is managed by the Meta Open Source bot.
#1797 opened Oct 4, 2025 by fegin Loading…
1
add option to disable ft checkpoints CLA Signed This label is managed by the Meta Open Source bot. fb-exported meta-exported
#1795 opened Oct 3, 2025 by tushar00jain Loading…
5
JointGraph-based Training Prototype CLA Signed This label is managed by the Meta Open Source bot.
#1794 opened Oct 3, 2025 by SherlockNoMad • Draft
[wip][simplefsdp] fix simplefsdp gradient_divide_factor CLA Signed This label is managed by the Meta Open Source bot.
#1793 opened Oct 3, 2025 by ruisizhang123 • Draft
Enable ROCm CI support ciflow/rocm CLA Signed This label is managed by the Meta Open Source bot. module: rocm
#1786 opened Oct 2, 2025 by akashveramd Loading…
@akashveramd
7
[debug don't merge] pr to reproduce error in aot_eager+post_grad_custom_post_pass CLA Signed This label is managed by the Meta Open Source bot.
#1785 opened Oct 2, 2025 by ruisizhang123 Loading…
Autoparallel DSv3 with local_map CLA Signed This label is managed by the Meta Open Source bot.
#1784 opened Oct 2, 2025 by xmfan • Draft
[RFC] Refactor attention and make attention mask an argument to the model CLA Signed This label is managed by the Meta Open Source bot.
#1776 opened Sep 30, 2025 by fegin Loading…
40
[ap] Knobs to enable reorder/bucketing/async_tp passes CLA Signed This label is managed by the Meta Open Source bot.
#1772 opened Sep 30, 2025 by IvanKobzarev Loading…
[mxfp8 MoE training] Support mxfp8 all to all in expert parallel CLA Signed This label is managed by the Meta Open Source bot.
#1765 opened Sep 26, 2025 by danielvegamyhre Loading…
Adding config options for deterministic execution CLA Signed This label is managed by the Meta Open Source bot.
#1761 opened Sep 26, 2025 by githubsgi Loading…
7
gpt-oss model enablement CLA Signed This label is managed by the Meta Open Source bot.
#1754 opened Sep 24, 2025 by wwwjn Loading…
5
improve profiler CLA Signed This label is managed by the Meta Open Source bot.
#1753 opened Sep 24, 2025 by tushar00jain • Draft
1
handle unable to load ft checkpoint CLA Signed This label is managed by the Meta Open Source bot.
#1752 opened Sep 24, 2025 by tushar00jain • Draft
1
[DONT REVIEW] Debug Async TP CI CLA Signed This label is managed by the Meta Open Source bot.
#1751 opened Sep 24, 2025 by fegin Loading…
Export MoE CLA Signed This label is managed by the Meta Open Source bot.
#1745 opened Sep 23, 2025 by kwen2501 • Draft
[DSV3] Replace 1D A2A with 2D A2A CLA Signed This label is managed by the Meta Open Source bot.
#1744 opened Sep 23, 2025 by kwen2501 • Draft
Add support for AC budget API CLA Signed This label is managed by the Meta Open Source bot.
#1731 opened Sep 21, 2025 by tohskai Loading…
14
[DONT REVIEW] debug ac(fsdp) in llama and deepseek CLA Signed This label is managed by the Meta Open Source bot.
#1724 opened Sep 18, 2025 by weifengpy • Draft
[Option 1] export example CLA Signed This label is managed by the Meta Open Source bot.
#1722 opened Sep 18, 2025 by H-Huang • Draft
[Option 2 Example] Dont land CLA Signed This label is managed by the Meta Open Source bot.
#1721 opened Sep 18, 2025 by H-Huang • Draft
1
[torchtitan][replicate] experimenting new replicate integration with torchtitan CLA Signed This label is managed by the Meta Open Source bot.
#1714 opened Sep 15, 2025 by anshul-si Loading…
27
[Do not merge] Reproduce AC(FSDP(moe.experts)) composibility issue CLA Signed This label is managed by the Meta Open Source bot.
#1709 opened Sep 12, 2025 by wwwjn • Draft
ProTip! Follow long discussions with comments:>50.