Skip to content

Pull requests: NVIDIA/Model-Optimizer

New pull request New
280 Open 1,671 Closed
280 Open 1,671 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 GPTQ Hessian coordinates after AWQ smoothing.
#2401 opened Sep 11, 2026 by sarahpardo • Draft
2
[OMNIML-5899] Add IQ1_S and IQ2_XS quantization and unified checkpoint export
#2381 opened Sep 11, 2026 by ChenhanYu Collaborator • Draft
3
Reuse a whole recipe via $import, and record which recipe reproduces each published NVIDIA checkpoint
#2376 opened Sep 11, 2026 by shengliangxu Collaborator • Draft
4
Add DASC recurrent state sparsity policy
#2375 opened Sep 10, 2026 by kaix-nv Contributor Loading…
49
Audit missing labels before release cherry-picks
#2373 opened Sep 10, 2026 by chadvoegele Contributor Loading…
7
Add Qwen 3.5 4B all-axis VLM campaign puzzletron_v2 Related to feature/puzzletron_v2 branch
#2372 opened Sep 10, 2026 by j-rausch Contributor Loading…
2
deprecate trtllm-build in weight_sparsity
#2371 opened Sep 10, 2026 by noeyy-mino Contributor Loading…
9
Invalidate sharded sparsity masks after updating the source mask
#2370 opened Sep 10, 2026 by MrCapricornLiu Loading…
2
Preserve zero batches in histogram calibration
#2369 opened Sep 10, 2026 by MrCapricornLiu Loading…
2
Release cached activations when exporting distillation models
#2368 opened Sep 10, 2026 by MrCapricornLiu Loading…
2
Fix folding transposed GPT-OSS and Llama 4 expert weights
#2367 opened Sep 10, 2026 by MrCapricornLiu Loading…
10
[Feature] Add Qwen3.8-27B local-Hessian PTQ recipes
#2363 opened Sep 9, 2026 by realAsma Contributor • Draft
2
Add structured Puzzletron results and progress puzzletron_v2 Related to feature/puzzletron_v2 branch
#2361 opened Sep 9, 2026 by j-rausch Contributor • 2/2 Loading…
8
Simplify Puzzletron configuration with recipes and sites puzzletron_v2 Related to feature/puzzletron_v2 branch
#2360 opened Sep 8, 2026 by j-rausch Contributor • 1/2 Loading…
17
Fix selected layers being skipped by post-init sparsity
#2345 opened Sep 6, 2026 by darkdi Loading…
2
fix: validate kd_loss_weight type and reject negative individual weights
#2343 opened Sep 5, 2026 by aryan-iconic Loading…
4
[Example] Qwen3.5 DSpark · E2E Training Example
#2340 opened Sep 5, 2026 by jinzex Loading…
10
[Fix] Calibrate non-decoder modules during layerwise quantization
#2339 opened Sep 4, 2026 by realAsma Contributor • Draft
70
[Example] [Frozen] Qwen3.5 DSpark · E2E Training Example
#2338 opened Sep 4, 2026 by jinzex • Draft
20
[5565357] Fix SDXL NVFP4 export and performance
#2336 opened Sep 4, 2026 by ajrasane Contributor Loading…
12
fix(quantization): fix nvfp4 availability check
#2331 opened Sep 4, 2026 by Lee-YNU Loading…
2
Add Parallel Decoding Distillation to FastGen
#2329 opened Sep 4, 2026 by mxinO Contributor Loading…
46
Rename modelopt_recipes/huggingface to model_type with backward-compat alias
#2328 opened Sep 3, 2026 by shengliangxu Collaborator • Draft
4
added new workflow puzzletron_v2 Related to feature/puzzletron_v2 branch
#2323 opened Sep 3, 2026 by grzegorz-k-karch Contributor Loading…
@grzegorz-k-karch
8
fp32 master weights for the DFlash draft, and keep them across a resume
#2322 opened Sep 3, 2026 by h-guo18 Contributor Loading…
6
ProTip! Add no:assignee to see everything that’s not assigned.