Skip to content

Pull requests: NVIDIA/TensorRT-Model-Optimizer

New pull request New
48 Open 248 Closed
48 Open 248 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 hf_quant_config with kv cache type
#557 opened Nov 14, 2025 by jenchen13 Loading…
3
Support Wan2.2 t2v diffusers quantization
#556 opened Nov 13, 2025 by shengliangxu Loading…
2
GPTQ Lite implementation
#555 opened Nov 13, 2025 by sugunav14 • Draft
2
[5455919] Fix Q/DQ/Cast placement in 'FP32 required' custom ops
#554 opened Nov 13, 2025 by gcunhase Loading…
4
Fix QLoRA example test
#553 opened Nov 13, 2025 by sugunav14 Loading…
@sugunav14
2
llama converter is self-contained now (no dependency on internal nvidia code)
#552 opened Nov 13, 2025 by danielkorzekwa Loading…
11
AutoQuantize minor improvement: limit grad enabled parameters, limit cpu-gpu sync during scoring
#551 opened Nov 13, 2025 by realAsma Loading…
1
Make all tensors on same device for svdquant with cpu-offloading
#550 opened Nov 13, 2025 by vishalpandya1990 Loading…
@vishalpandya1990
1
Feat: Eagle3 HF Online - support nemotron models
#548 opened Nov 13, 2025 by h-guo18 Loading…
6
[5336870] AutoCast: Unblock LSTM from conversion
#544 opened Nov 12, 2025 by galagam Loading…
@galagam
1
[5643020] AutoCast use onnxscript for opset conversion
#542 opened Nov 12, 2025 by galagam Loading…
@galagam
1
Specdec Bench: vLLM reqid, SGL path, conc > 1 metric fix
#541 opened Nov 12, 2025 by IzzyPutterman Loading…
3
[OMNIML-3015]Add per tensor/per channel MSE calibrator
#540 opened Nov 12, 2025 by Fridah-nv Loading…
2 tasks
@Fridah-nv
1
[OMNIML-2850] [3/n] Adds sparse attention calibration
#538 opened Nov 11, 2025 by kaix-nv Loading…
1
Use ONNX DQ node instead of DQ custom-op for activation dequantization in nvfp4
#536 opened Nov 11, 2025 by vishalpandya1990 Loading…
1
Optimize NVFP4 Triton kernel
#533 opened Nov 11, 2025 by mxinO • Draft
@mxinO
2
[OMNIML-2852] [2/n] Add Core Sparse Attention Infrastructure
#527 opened Nov 7, 2025 by kaix-nv Loading…
1
parallel eagle draft
#523 opened Nov 6, 2025 by yeyu-nvidia • Draft
2
[Bug #193] fix fp8 blockwise real quantization
#522 opened Nov 6, 2025 by meenchen Loading…
@meenchen
1
Support AWQ fake quant for vLLM MoE models
#521 opened Nov 6, 2025 by meenchen • Draft
@meenchen
1
Fix BMM style MoE export in fp8_pc_pt recipe
#515 opened Nov 5, 2025 by Edwardf0t1 Loading…
5
[Draft] [5526696] Add kv cache quantization support for onnx quantization
#486 opened Oct 31, 2025 by zhanghaoc Loading…
@zhanghaoc
19
Yeyu/set block
#480 opened Oct 28, 2025 by yeyu-nvidia • Draft
1
feat: add onnxslim support
#478 opened Oct 28, 2025 by inisis Loading…
10
Feat: Eagle3 HF Online - support nemotron models
#463 opened Oct 25, 2025 by h-guo18 Loading…
@h-guo18
24
ProTip! Adding no:label will show everything without a label.