Skip to content

Pull requests: NVIDIA-NeMo/Automodel

New pull request New
187 Open 2,747 Closed
187 Open 2,747 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

fix(vlm): truncate only the token axis in gemma4 collate
#3531 opened Aug 13, 2026 by athitten Contributor • Draft
2 of 3 tasks
1
fix(pp): keep accumulated grads across stage re-initialization
#3530 opened Aug 13, 2026 by athitten Contributor • Draft
2 of 3 tasks
1
test(pp): add gemma4 parallelism parity tests
#3529 opened Aug 13, 2026 by athitten Contributor • Draft
2 of 3 tasks
2
fix(packaging): gate FlashOptim to Linux community-request
#3528 opened Aug 13, 2026 by ting-hong-shieh • Draft
4 tasks done
2
refactor(vlm): resolve packing token counts from the processor
#3527 opened Aug 12, 2026 by HuiyingLi Contributor • Draft
3 tasks done
1
feat(loss): return per-depth MTP losses
#3525 opened Aug 12, 2026 by Slyne Contributor Loading…
3 tasks done
6
fix(moe): stabilize async DeepEP checkpointing
#3522 opened Aug 12, 2026 by pzelasko Contributor Loading…
3
refactor(models): group model packages under their transformers/diffusers bridge
#3521 opened Aug 12, 2026 by akoumpa Contributor • Draft
1
ci: Update transformers to latest version 5.15.0
#3516 opened Aug 12, 2026 by svcnvidia-nemo-ci Contributor Loading…
6
feat(datasets): Datum record and collate-time THD sequence packing
#3514 opened Aug 12, 2026 by HuiyingLi Contributor • Draft
5
feat(gemma4): add E-series tensor parallelism
#3512 opened Aug 12, 2026 by HuiyingLi Contributor Loading…
1
fix(minimax_m3_vl): declare and honour logits_to_keep so fused losses are kept community-request
#3511 opened Aug 12, 2026 by coderaBruce Loading…
1
15
feat(wan): wan animate 2
#3496 opened Aug 11, 2026 by zyzhou5 Contributor Loading…
42
ci: validate dev ephemeral runner routing (do not merge)
#3494 opened Aug 11, 2026 by chtruong814 Contributor Loading…
6 tasks done
5
ci: add GPU runner label suffix switch
#3486 opened Aug 11, 2026 by chtruong814 Contributor Loading…
6 tasks done
7
build(ci): harden repository test image
#3484 opened Aug 10, 2026 by ko3n1g Contributor • Draft
3 tasks done
1
perf(lora): reduce tensor parallel redistribution
#3477 opened Aug 10, 2026 by akoumpa Contributor • Draft
20
feat(datasets): add a token-budget dynamic batch sampler community-request waiting-on-maintainers Waiting on maintainers to respond
#3473 opened Aug 10, 2026 by khazic Contributor Loading…
5
fix(vlm): use FP32 master weights for Nemotron Omni and other fixes
#3444 opened Aug 6, 2026 by HuiyingLi Contributor • Draft
3
fix(kimi_k3): apply the HF renames to expert LoRA keys on save community-request
#3435 opened Aug 5, 2026 by stanley1208 Contributor Loading…
1
11
feat(kernels): lean Hub integration via Transformers passthrough community-request waiting-on-customer Waiting on the original author to respond
#3426 opened Aug 5, 2026 by kashif Contributor Loading…
1 of 2 tasks
165
fix(checkpoint): shard extracted exports by size
#3423 opened Aug 5, 2026 by oliverholworthy Contributor • Draft
3 tasks done
1
fix(model_init): stream Devstral FP8 checkpoints
#3413 opened Aug 4, 2026 by yuhezhang-ai Contributor Loading…
1
9
perf(moe): add HybridEP compact routing
#3404 opened Aug 4, 2026 by akoumpa Contributor • Draft
3
test: mirror unit test layout with source packages
#3388 opened Aug 4, 2026 by akoumpa Contributor • Draft
3
ProTip! Updated in the last three days: updated:>2026-08-09.