Hybrid-Sensitivity-Weighted-Quantization (HSWQ)
High-fidelity ConvRot INT8 reverse hybrid quantization for SDXL diffusion models. HSWQ uses per-layer trajectory-impact measurement instead of naive uniform cast, converting only the K lowest-impact layers to ConvRot INT8 while keeping every other layer at FP16. This is highly useful for users who need to strictly manage their VRAM resources while maintaining maximum image quality.
📦 Available Models
Filename convention: <model>_hswq_1on_re<K>_convrot_int8.safetensors — reverse hybrid with K lowest-impact layers converted to ConvRot INT8, bias correction ON (1on), everything else FP16.
| Filename | Base Model | Version | License |
|---|---|---|---|
JANKUTrainedChenkinNoobai_v777_hswq_1on_re550_convrot_int8.safetensors |
JANKU Trained Chenkin & Noobai-Rouwei (Illustrious-XL) | v777 | Fair AI Public License 1.0-SD |
bluePencilXL_v031_hswq_1on_re570_convrot_int8.safetensors |
blue_pencil-XL | v0.3.1 | CreativeML Open RAIL++-M |
epicrealismXL_pureFix_hswq_1on_re570_convrot_int8.safetensors |
epiCRealism XL | pureFix | CreativeML Open RAIL++-M |
koronemixIllustrious_v70_hswq_1on_re550_convrot_int8.safetensors |
koronemixIllustrious | v70 | Fair AI Public License 1.0-SD |
koronemixVpred_v20_hswq_1on_re550_convrot_int8.safetensors |
koronemixVpred | v2.0 | CreativeML Open RAIL++-M |
novaAnimeXL_ilV190_hswq_1on_re599_convrot_int8.safetensors |
Nova Anime XL | ilV190 | Fair AI Public License 1.0-SD |
novaAsianXL_illustriousV70_hswq_1on_re550_convrot_int8.safetensors |
Nova Asian XL | v7.0 | Fair AI Public License 1.0-SD |
oneObsession_v24_hswq_1on_re572_convrot_int8.safetensors |
OneObsession | v24 | CreativeML Open RAIL++-M |
prefectIllustriousXL_v8_hswq_1on_re610_convrot_int8.safetensors |
Prefect Illustrious XL | v8 | Fair AI Public License 1.0-SD |
realvisxlV30_v30TurboBakedvae_hswq_1on_re650_convrot_int8.safetensors |
RealVisXL V3.0 (Turbo) | v3.0 Turbo | CreativeML Open RAIL++-M |
realvisxlV50_v40Bakedvae_hswq_1on_re550_convrot_int8.safetensors |
RealVisXL V5.0 (Lightning) | v4.0 BakedVAE | CreativeML Open RAIL++-M |
realvisxlV50_v50Bakedvae_hswq_1on_re550_convrot_int8.safetensors |
RealVisXL V5.0 (Lightning) | v5.0 BakedVAE | CreativeML Open RAIL++-M |
unholyDesireMixSinister_v90_hswq_1on_re590_convrot_int8.safetensors |
Unholy Desire Mix - Sinister | v9.0 | Fair AI Public License 1.0-SD |
uwazumimixILL_v50_hswq_1on_re720_convrot_int8.safetensors |
UwazumiMix | v5.0 | Fair AI Public License 1.0-SD |
waiANIPONYXL_v140_hswq_1on_re650_convrot_int8.safetensors |
WAI-ANI-PONY-XL | v14.0 | Fair AI Public License 1.0-SD |
waiANIPONYXL_v90_hswq_1on_re650_convrot_int8.safetensors |
WAI-ANI-PONY-XL | v9.0 | Fair AI Public License 1.0-SD |
waiIllustriousSDXL_v170_hswq_1on_re597_convrot_int8.safetensors |
Illustrious-XL v1.7 (WAI-illustrious-SDXL) | v17.0 | Fair AI Public License 1.0-SD |
waiREALCN_v150_hswq_1on_re630_convrot_int8.safetensors |
WAI-REAL_CN | v15.0 | Fair AI Public License 1.0-SD |
waiREALISM_v10_hswq_1on_re590_convrot_int8.safetensors |
WAI-REALISM | v1.0 | 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.
- JANKU Trained Chenkin & Noobai-Rouwei (Illustrious-XL): Created by janxd.
- blue_pencil-XL: Created by Euge_us.
- epiCRealism XL: Created by epinikion.
- WAI-illustrious-SDXL / WAI-REAL_CN / WAI-REALISM / WAI-ANI-PONY-XL: Created by WAI0731.
- koronemixIllustrious / koronemixVpred: Created by koronen.
- Nova Anime XL / Nova Asian XL: Created by Crody.
- Prefect Illustrious XL: Created by Goofy_Ai.
- OneObsession: Created by Polyhedron.
- RealVisXL: Created by SG_161222.
- Unholy Desire Mix - Sinister: Created by UnholyDesiresStudio.
- UwazumiMix: Created by UWAZUMI.
Disclaimer: These models are provided for optimization and research purposes. Please adhere to the original licenses of the base models.
📚 Detailed documentation
- md/How to quantize SDXL.md
- md/Diag_Reverse_SDXL_v1.1_Technical_Guide.md
- benchmark result/benchmark_sdxl_int8.md
🔒 Security
Backdoor audit. Each file's audit status - verdict strength, weight coverage, the record behind it, and the limits of what that measurement can show - is in Backdoor Audit Verdicts. The audit code, the procedure and the raw records are in the same repository.