Hybrid-Sensitivity-Weighted-Quantization (HSWQ)
High-fidelity ConvRot NVFP4 quantization for diffusion models (SDXL). HSWQ uses sensitivity and importance analysis instead of naive uniform cast. This is highly useful for users who need to strictly manage their VRAM resources while maintaining maximum image quality.
ComfyUI Load Diffusion Model nvfp4 pack with FULL ConvRot (Linear→NVFP4, Conv2d→INT8 int8_tensorwise) after DualMonitor + V4 pack-MSE FP16 protection under a fixed 600 MiB budget. Keep ratio is 0 (r0); calib writes NVFP4 .input_scale. SDXL pack scripts: hswq_convert_nvfp4_1.0.py (HSWQ) and native_convert_nvfp4.py (native).
Technical details: https://github.com/ussoewwin/Hybrid-Sensitivity-Weighted-Quantization
How to quantize (SDXL ConvRot NVFP4): md/How to quantize SDXL NVFP4.md
ComfyUI Loader for ConvRot NVFP4: To use these models in ComfyUI, please use this custom node: ComfyUI-HSWQ-Loader-and-Tools
Post-convert fidelity bench (integrated, default ON): After save, hswq_convert_nvfp4_1.0.py and native_convert_nvfp4.py clear parent VRAM, then automatically run benchmark/nvfp4bench_sdxl.py with --fp16 = the FP16 input, --nvfp4 = the saved pack, and a fixed --prompt / --seed (not inventable parent CLI overrides). Pass --no-bench to skip. Standalone re-runs use the same nvfp4bench_sdxl.py command shape as in the How-to.
SDXL ConvRot NVFP4 Benchmark Test Results (published tables): test/benchmark_convrotnvfp4.md
Benchmark (Reference)
| Model | SSIM (Avg) | File size | Compatibility |
|---|---|---|---|
| Original FP16 | 1.0000 | 100% | High |
| HSWQ ConvRot NVFP4 | 0.92-0.98 | 60% (FP16 mixed) | High (ComfyUI NVFP4) |
📦 Available Models
| Filename | Base Model | Version | License |
|---|---|---|---|
waiIllustriousSDXL_v170_hswq_r32_convrot_nvfp4.safetensors |
Illustrious-XL v1.7 (WAI-illustrious-SDXL) | v17.0 (HF weight) | Fair AI Public License 1.0-SD |
JANKUTrainedChenkinNoobai_v777_hswq_r32_+200_nvfp4.safetensors |
JANKU Trained Chenkin & Noobai-Rouwei (Illustrious-XL) | v777 | Fair AI Public License 1.0-SD |
animemix_v80_hswq_r32_nvfp4.safetensors |
AnimeMix | v8.0 | Fair AI Public License 1.0-SD |
koronemixIllustrious_v70_hswq_r32_convrot_nvfp4.safetensors |
koronemixIllustrious | v70 | Fair AI Public License 1.0-SD |
📜 Credits & License
🏆 Special Acknowledgement
We extend our deepest respect and gratitude to the Nunchaku Team for their groundbreaking work on SVDQ quantization and for sharing their models with the community. This collection relies heavily on their research and original implementation.
- Original Repository: nunchaku-tech/nunchaku-sdxl
Base Models
These models are derivatives of their respective creators. All credit for aesthetic tuning and model training belongs to the original creators.
- WAI-illustrious-SDXL: Created by WAI0731.
- JANKU Trained Chenkin & Noobai-Rouwei (Illustrious-XL): Created by janxd.
- AnimeMix / koronemixIllustrious: Created by koronen.
Disclaimer: These models are provided for optimization and research purposes. Please adhere to the original licenses of the base models.