nvidia-modelopt is a unified library of state-of-the-art model optimization techniques like quantization, pruning, distillation, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorRT-LLM or TensorRT to optimize inference speed.
active 2025-03-28 → 2025-12-07 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 255 days from 2025-03-28 to 2025-12-07. Pushes: 1,535 total, peak 44 in a day. Pull requests: 244 total, peak 9 in a day. Issues: 292 total, peak 69 in a day. Comments: 1,433 total, peak 72 in a day. Stars: 379 total, peak 27 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| coderabbitai[bot] | 707 | 1 | 1 | 375 |
| copy-pr-bot[bot] | 580 | 515 | 1 | 64 |
| kevalmorabia97 | 487 | 175 | 65 | 137 |
| cjluo-nv | 265 | 34 | 10 | 75 |
| realAsma | 262 | 75 | 10 | 81 |
| danielkorzekwa | 212 | 143 | 6 | 30 |
| ajrasane | 203 | 52 | 15 | 27 |
| yeyu-nvidia | 144 | 44 | 3 | 50 |
| h-guo18 | 133 | 65 | 10 | 28 |
| i-riyad | 128 | 43 | 8 | 46 |
| jingyu-ml | 120 | 58 | 2 | 31 |
| codecov[bot] | 119 | 0 | 0 | 119 |
| Edwardf0t1 | 91 | 35 | 9 | 22 |
| meenchen | 73 | 13 | 3 | 27 |
| kaix-nv | 69 | 63 | 5 | 0 |
| AAnoosheh | 65 | 16 | 14 | 14 |
| ChenhanYu | 57 | 12 | 3 | 15 |
| shengliangxu | 56 | 30 | 1 | 17 |
| jenchen13 | 53 | 31 | 6 | 8 |
| gcunhase | 51 | 7 | 7 | 20 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#622kevalmorabia972025-12-07 06:52Updated dependencies in Windows examples
- Issue comment#614Hyubo2025-12-06 08:39ONNX export failure with mtq.FP8_DEFAULT_CFG and mtq.NVFP4_DEFAULT_CFG
- Issue#658PonyPinkPie2025-12-06 08:17Theoretically, can inputs and outputs data type of Add opt be FP8 mode?
- Pull request#654h-guo182025-12-06 01:00
- Issue comment#657copy-pr-bot[bot]2025-12-06 00:30config file based modelopt config 1/N
- Issue comment#500shivghai2025-12-05 21:38activation_scaling_factor missing in TensorRT-LLM Checkpoint FP8 Export
- Issue comment#654copy-pr-bot[bot]2025-12-05 19:51Fix: update file path in eagle3 scripts
- Issue comment#653codecov[bot]2025-12-05 19:47Support model export for int4 wo
- Pull request#653meenchen2025-12-05 19:36
- Issue comment#636gcunhase2025-12-05 14:40[OMNIML-2244] Create the nvfp4 quant exporter
- Issue comment#650codecov[bot]2025-12-05 08:51Noeyy/add new ckpts test cases
- Pull request#651sugunav142025-12-05 08:51
- Pull request#651sugunav142025-12-05 08:51
- Issue#88github-actions[bot]2025-12-05 03:32Can you quantify the Multi-scale Deformable Attention module?
- Issue comment#88github-actions[bot]2025-12-05 03:32Can you quantify the Multi-scale Deformable Attention module?
- Issue#328github-actions[bot]2025-12-05 03:32Bug when quantize GLM-MoE model
- Issue#528github-actions[bot]2025-12-05 03:32Quantize DeepSeek R1 BUG
- Issue comment#649codecov[bot]2025-12-04 23:03[NVBUG: 5701937]Clear GPU cache for 3D weight tensors
- Issue#54realAsma2025-12-04 22:56How to implement LSQ using pytorch_quant
- Issue comment#54realAsma2025-12-04 22:56How to implement LSQ using pytorch_quant
- Pull request#649cjluo-nv2025-12-04 22:53
- Issue#595realAsma2025-12-04 22:45Will the NVFP4_AFFINE_KV_CFG with static bias config take effect?
- Issue#595realAsma2025-12-04 22:45Will the NVFP4_AFFINE_KV_CFG with static bias config take effect?
- Issue#647realAsma2025-12-04 22:29Support PTQ for Qwen3 Omni
- Issue#647realAsma2025-12-04 22:26Support PTQ for Qwen3 Omni
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 379 stars here means stars gained during the window, not the repo's star count.