Skip to content

Pull requests: NVIDIA/Model-Optimizer

New pull request New
199 Open 1,192 Closed
199 Open 1,192 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

[Chore]: Update Dflash recipes to use dpace
#1775 opened Jun 20, 2026 by h-guo18 Contributor • Draft
4
[OMNIML-5091] specdec_bench cell t0_d7 — nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4 / MTP / trtllm
#1774 opened Jun 19, 2026 by ChenhanYu Collaborator • Draft
3
Experimental claude skill for puzzletron algoritgm
#1769 opened Jun 18, 2026 by danielkorzekwa Contributor Loading…
7
feat(launcher): add Megatron-Bridge quantize/generate/export wrappers
#1767 opened Jun 17, 2026 by yueshen2016 Contributor Loading…
20
launcher: package as modelopt_launcher; mcp: call console script directly
#1766 opened Jun 17, 2026 by ChenhanYu Collaborator Loading…
5 tasks
@ChenhanYu
8
fix(puzzletron): correct val_dataset_name from 'valid' to 'validation'
#1765 opened Jun 17, 2026 by TheSabari07 Contributor Loading…
2
feat(recipes): add nvfp4_mlp_only-kv_fp8-novit (exclude VL vision tower)
#1760 opened Jun 17, 2026 by Edwardf0t1 Contributor Loading…
4
refactor(examples): rename llm_ptq → hf_ptq (symlink for back-compat)
#1759 opened Jun 17, 2026 by Edwardf0t1 Contributor Loading…
16
Add p quantization to our triton fa kernel
#1757 opened Jun 16, 2026 by sychen52 Contributor Loading…
@sychen52
15
[OMNIML-5003] Support non-gated fused MoE experts (NemotronH) in HF PTQ
#1756 opened Jun 16, 2026 by jenchen13 Contributor Loading…
10
DFlash for MiniMax-M3 (WIP): synthesis thinking-mode mix
#1749 opened Jun 16, 2026 by yeyu-nvidia Contributor • Draft
3
[minor] Pre-build CUDA quantization extensions before unit tests run
#1748 opened Jun 16, 2026 by shengliangxu Collaborator Loading…
9
Add Minitron pruning support for GatedDeltaNet, MLA, and latent MoE
#1747 opened Jun 16, 2026 by kevalmorabia97 Collaborator Loading…
30
fix(autocast): add missing guard for use_standalone_type_inference
#1743 opened Jun 15, 2026 by agkphysics Loading…
2
[OMNIML-5024] cell_t0_d3
#1733 opened Jun 15, 2026 by ChenhanYu Collaborator • Draft
2
fix(megatron): handle MambaMixer conv1d refactor in importer/exporter
#1730 opened Jun 15, 2026 by AAnoosheh Contributor Loading…
2 tasks done
@AAnoosheh
3
[DRAFT] Add heterogeneous AnyModel distillation example for Puzzletron.
#1725 opened Jun 15, 2026 by chochowski Contributor Loading…
1
11
Fix autotune warm restart to retry failed schemes
#1722 opened Jun 15, 2026 by willg-nv Contributor Loading…
2
[OMNIML-5091] specdec_bench cell t0_d7 — nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4 / MTP / trtllm
#1721 opened Jun 14, 2026 by ChenhanYu Collaborator • Draft
3
[Quantization] Support NVFP4 for inline-swiglu fused MoE experts (MiniMax-M3 / MiniMaxM3VLExperts)
#1719 opened Jun 14, 2026 by yifjiang • Draft
4 of 5 tasks
2
[OMNIML-5072] Triton fakequant adapter for N-quantizer per-expert path
#1717 opened Jun 14, 2026 by hychiang-git Contributor • Draft
2
Support Mamba direct conv1d quant export
#1716 opened Jun 14, 2026 by mxinO Contributor • Draft
5
[OMNIML-5100] specdec_bench cell t1_d7 — nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 / MTP / vllm
#1715 opened Jun 14, 2026 by ChenhanYu Collaborator • Draft
3
[OMNIML-5097] cell_t0_d3
#1713 opened Jun 14, 2026 by ChenhanYu Collaborator • Draft
2
fix(quantization): detect fused MoE experts without act_fn (MiniMax-M3)
#1711 opened Jun 14, 2026 by Edwardf0t1 Contributor Loading…
3
ProTip! Filter pull requests by the default branch with base:main.