Skip to content

Pull requests: pytorch/torchtitan

New pull request New
50 Open 580 Closed
50 Open 580 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

Adding a simple optimizer registry. CLA Signed This label is managed by the Meta Open Source bot.
#876 opened Feb 21, 2025 by balancap Loading…
2
Create logging directory in wandb logger CLA Signed This label is managed by the Meta Open Source bot.
#874 opened Feb 21, 2025 by K-H-Ismail Loading…
Refactor Checkpointer CLA Signed This label is managed by the Meta Open Source bot.
#871 opened Feb 20, 2025 by fegin Loading…
Show GC execution time CLA Signed This label is managed by the Meta Open Source bot.
#870 opened Feb 20, 2025 by fegin Loading…
Configure arbitrary frozen modules via config CLA Signed This label is managed by the Meta Open Source bot.
#869 opened Feb 20, 2025 by lkhphuc Loading…
3
add structure to torchtitan files CLA Signed This label is managed by the Meta Open Source bot.
#867 opened Feb 20, 2025 by tianyu-l Loading…
11
Do full GC for checkpointing related GC calls CLA Signed This label is managed by the Meta Open Source bot.
#858 opened Feb 19, 2025 by fegin • Draft
1
[Reland] Add Dynamic Model Import and ModelSpec Definition (#837) ci-no-td CLA Signed This label is managed by the Meta Open Source bot.
#854 opened Feb 18, 2025 by fduwjj Loading…
[Not for landing] piggy back on titan for scale init test CLA Signed This label is managed by the Meta Open Source bot.
#841 opened Feb 13, 2025 by fduwjj • Draft
2
Avoid using skip() in hf_datasets CLA Signed This label is managed by the Meta Open Source bot.
#835 opened Feb 12, 2025 by fegin • Draft
4
[NOT READY TO LAND] Integrate TorchFT CLA Signed This label is managed by the Meta Open Source bot.
#834 opened Feb 11, 2025 by fegin Loading…
Add force_recompute_fp8_weight_in_bwd when FSDP CLA Signed This label is managed by the Meta Open Source bot.
#832 opened Feb 11, 2025 by c0g Loading…
2
Intial add of DeepSeek-V3 model CLA Signed This label is managed by the Meta Open Source bot.
#830 opened Feb 11, 2025 by kwen2501 • Draft
1
profile with modules and stack CLA Signed This label is managed by the Meta Open Source bot.
#829 opened Feb 10, 2025 by carmocca Loading…
4
[not for land] add repro script for https://github.com/pytorch/torchtitan/pull/808 CLA Signed This label is managed by the Meta Open Source bot.
#815 opened Feb 1, 2025 by danielvegamyhre • Draft
add configuration for float8 with rowwise scaling, via recipe lookup CLA Signed This label is managed by the Meta Open Source bot.
#808 opened Jan 27, 2025 by vkuzo Loading…
35
[WIP][RFC] TorchFT integration CLA Signed This label is managed by the Meta Open Source bot.
#806 opened Jan 27, 2025 by fegin • Draft
[cp] Add cudnn attention support to Context Parallel CLA Signed This label is managed by the Meta Open Source bot.
#796 opened Jan 17, 2025 by XilunWu • Draft
[debug only] microbatching CLA Signed This label is managed by the Meta Open Source bot.
#795 opened Jan 16, 2025 by tianyu-l • Draft
Make CheckpointManager friendlier to custom StorageWriter/StorageReader CLA Signed This label is managed by the Meta Open Source bot.
#789 opened Jan 12, 2025 by dimdi-y Loading…
6
Register backward hook for the whole optim_dict to enable working at multi schedule pp CLA Signed This label is managed by the Meta Open Source bot.
#780 opened Jan 7, 2025 by mori360 • Draft
[Not for land] Integrate float8nocompile, an experimental feature for high performance CLA Signed This label is managed by the Meta Open Source bot.
#778 opened Jan 7, 2025 by danielvegamyhre Loading…
8
[PoC] Typed JobConfig CLA Signed This label is managed by the Meta Open Source bot.
#767 opened Jan 1, 2025 by jaysonfrancis Loading…
1
[MoE][PoC] Expert Parallel: dp2ep CLA Signed This label is managed by the Meta Open Source bot.
#732 opened Dec 12, 2024 by tianyu-l • Draft
4
[MoE][PoC] Expert Parallel: tp and tp2ep CLA Signed This label is managed by the Meta Open Source bot.
#731 opened Dec 12, 2024 by tianyu-l • Draft
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.