Skip to content

Pull requests: NVIDIA-NeMo/Automodel

New pull request New
178 Open 2,721 Closed
178 Open 2,721 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(models): preserve lm-head dtype boundaries r0.6.0 Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#3491 opened Aug 11, 2026 by jQizhang Contributor Loading…
3 tasks done
2
ci: add GPU runner label suffix switch
#3486 opened Aug 11, 2026 by chtruong814 Contributor Loading…
6 tasks done
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
#3473 opened Aug 10, 2026 by khazic Contributor Loading…
5
fix(deepseek_v4): avoid TileLang boolx8 backward codegen
#3467 opened Aug 9, 2026 by HuiyingLi Contributor • Draft
3 tasks done
1
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 waiting-on-maintainers Waiting on maintainers to respond
#3435 opened Aug 5, 2026 by stanley1208 Contributor Loading…
1
7
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
feat(moe): integrate Mixture-of-Kittens backend
#3422 opened Aug 5, 2026 by HuiyingLi Contributor Loading…
1 of 3 tasks
1
6
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
feat(datasets): support SynthTraces agent SFT community-request
#3371 opened Aug 3, 2026 by khazic Contributor Loading…
1
61
feat(pretrain): add validated Qwen3-8B recipes
#3368 opened Aug 2, 2026 by akoumpa Contributor • Draft
3 tasks done
1
perf: restore Qwen3 Next HybridEP throughput r0.6.0 Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#3367 opened Aug 2, 2026 by akoumpa Contributor • Draft
17
feat(speculative): add Ulysses context parallelism for the EAGLE-3 draft community-request waiting-on-customer Waiting on the original author to respond
#3356 opened Aug 1, 2026 by khazic Contributor Loading…
5
feat(dllm): add SCDD training and sampling community-request
#3354 opened Aug 1, 2026 by kashif Contributor Loading…
21
feat(training): add weighted domain mixtures
#3303 opened Jul 30, 2026 by akoumpa Contributor Loading…
3 tasks done
28
feat(diffusion): integrate DMD2 into the native trainer
#3298 opened Jul 30, 2026 by jingyu-ml Loading…
26
feat(distributed): migrate Megatron FSDP to MCore v2
#3292 opened Jul 29, 2026 by wujingyue • Draft
7
chore: update claude review to FW-CI-templates v1.8.5
#3291 opened Jul 29, 2026 by kajalj22 Contributor Loading…
1
feat(speculative): add Kimi K3 DFlash draft training community-request waiting-on-customer Waiting on the original author to respond
#3287 opened Jul 29, 2026 by khazic Contributor Loading…
7
fix(peft): make fallback LoRA selective-AC safe r0.6.0 Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#3274 opened Jul 28, 2026 by HuiyingLi Contributor • Draft
3 tasks done
3
ProTip! Filter pull requests by the default branch with base:main.