Skip to content

Pull requests: PaddlePaddle/PaddleFormers

New pull request New
192 Open 2,844 Closed
192 Open 2,844 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

minicpm support contributor
#4421 opened May 9, 2026 by fjaijf Loading…
2 tasks done
2
fix separate_mtp_headloss aoa for hf save & load
#4420 opened May 9, 2026 by Wennie396 Contributor Loading…
2
Fix reshape for zcc master weight save
#4419 opened May 9, 2026 by changeyoung98 Contributor Loading…
2 tasks
2
迁移Benchmark的case至场内代码库
#4418 opened May 9, 2026 by XieYunshen Contributor Loading…
2 tasks
1
feat:internvl test contributor
#4416 opened May 9, 2026 by fjaijf Loading…
2 tasks done
2
Add empty skills directory
#4415 opened May 9, 2026 by a31413510 Collaborator Loading…
1 task
1
fix:minicpm_test contributor
#4410 opened May 8, 2026 by fjaijf Loading…
2 tasks done
1
Add InternVL support contributor
#4408 opened May 8, 2026 by fjaijf Loading…
2 tasks done
1
[Qwen2MoE] Fix h1_normed gradient accumulation order divergence via probe hook
#4407 opened May 8, 2026 by a31413510 Collaborator Loading…
1
Add Molmo model support
#4406 opened May 8, 2026 by yicycyc Loading…
2
[Qwen2MoE] Fix gradient alignment: remove fake_path and use clear_grad(set_to_zero=False)
#4405 opened May 8, 2026 by a31413510 Collaborator Loading…
1
fix SFTDataset : refine num_proc and num_workers co-existence check
#4404 opened May 8, 2026 by weiyixuanxx Contributor Loading…
2 tasks done
2
fix(qwen2_moe): use active-expert-only iteration in SparseMoeBlock
#4400 opened May 7, 2026 by a31413510 Collaborator Loading…
1
test formers bot
#4386 opened Apr 29, 2026 by zjjlivein Collaborator Loading…
2 tasks
5
Feature/add model openelm contributor
#4372 opened Apr 28, 2026 by learncat163 Loading…
4
[Qwen3Moe] Add pad_token_id fallback
#4371 opened Apr 28, 2026 by SigureMo Member Loading…
5
[fix] new epoch acc loss logic fix
#4365 opened Apr 27, 2026 by wacxr123 Contributor Loading…
2
Add granite
#4364 opened Apr 27, 2026 by Minestar6 Loading…
3
[Trainer | Cherry-Pick] feat(callback): support Fleet MoE class and GlobalRNGCallback
#4359 opened Apr 27, 2026 by hushenwei2000 Contributor Loading…
3
[Qwen3MoE] Fix gradient alignment between PaddlePaddle and PyTorch/HF
#4358 opened Apr 25, 2026 by a31413510 Collaborator Loading…
1
Add auto-subabtch config
#4355 opened Apr 24, 2026 by Difers Contributor Loading…
2 tasks
4
[release/1.1] Add high_precision_rope cfg contributor
#4354 opened Apr 24, 2026 by risemeup1111 Loading…
2 tasks
2
add olmo2 model
#4353 opened Apr 24, 2026 by yicycyc Loading…
7
fix aoa shared_head contributor
#4352 opened Apr 24, 2026 by Lcysabcu Collaborator Loading…
2 tasks
2
[Deps] pin paddlecodec to >=0.1, <0.2 for Paddle 3.3 compatibility
#4349 opened Apr 24, 2026 by SigureMo Member Loading…
1 of 2 tasks
4
ProTip! Filter pull requests by the default branch with base:develop.