-
Notifications
You must be signed in to change notification settings - Fork 513
All issues
Issue creation is restricted in this repository
- #1699 · Trenton-Starkey opened
on Jun 12, 2026 4
Issues
is:issue state:open
is:issue state:open
Search results
[ONNX PTQ] Support for third-party custom ORT/TRT plugins in static calibration
bugSomething isn't workingSomething isn't workingStatus: Open.#2016 In NVIDIA/Model-Optimizer;Recommended FP8 recipe for Blackwell (B200)? Per-tensor
FP8_DEFAULT_CFGunderperforms BF16 at prefill; NVFP4 much faster. Also: exported FP8 checkpoint has no KVq/k/v_scalebugSomething isn't workingSomething isn't workingStatus: Open.#2015 In NVIDIA/Model-Optimizer;- Status: Open.#2011 In NVIDIA/Model-Optimizer;
- Status: Open.#2002 In NVIDIA/Model-Optimizer;
- Status: Open.#2001 In NVIDIA/Model-Optimizer;
Nvidia Modelopt structured 2:4 weight sparsity showing no speed improvement compared to dense model
bugSomething isn't workingSomething isn't workingStatus: Open.#1974 In NVIDIA/Model-Optimizer;ModelOpt CI is failing with: Windows fatal exception
bugSomething isn't workingSomething isn't workingStatus: Open.#1969 In NVIDIA/Model-Optimizer;Quantizing error for Qwen3.6 35B A3B
bugSomething isn't workingSomething isn't workingStatus: Open.#1933 In NVIDIA/Model-Optimizer;- Status: Open.#1926 In NVIDIA/Model-Optimizer;
- Status: Open.#1902 In NVIDIA/Model-Optimizer;
warnings.warn(f"RealQuantLinear: No real-quant GEMM found: {self}.")
bugSomething isn't workingSomething isn't workingStatus: Open.#1895 In NVIDIA/Model-Optimizer;NVFP4 quantized (shared) lm_head fails fake-quant in training forward: ScaledE4M3Function 'Only support E=4 & M=3'
bugSomething isn't workingSomething isn't workingstaleNot updated in a long timeNot updated in a long timeStatus: Open.#1875 In NVIDIA/Model-Optimizer;