pipeline_tag: text-to-image
tags:
- Flux1
- Chroma
- Krea
- SVDQaunt
- Project0
- SVDQ
- Quantization
language:
- en
base_model_relation: quantized
library_name: diffusers
This repository contains Nunchaku-quantized (SVDQ) versions of , a text-to-image model based on Flux Dev and Flux Krea created by
If you like the model, please consider liking, reviewing and tipping the creator.
Model Files
svdq-int4_r32-Project0-REAL1SM_V3.safetensors: SVDQuant INT4 (rank 32) CenKreChro 1.0 model. For users with non-Blackwell GPUs (pre-50-series).
svdq-fp4_r32-Project0-REAL1SM_V3.safetensors: SVDQuant NVFP4 (rank 32) CenKreChro 1.0 model. For users with Blackwell GPUs (50-series).
svdq-int4_r32-CenKreChro.safetensors: SVDQuant INT4 (rank 32) CenKreChro 1.0 model. For users with non-Blackwell GPUs (pre-50-series).
svdq-fp4_r32-CenKreChro.safetensors: SVDQuant NVFP4 (rank 32) CenKreChro 1.0 model. For users with Blackwell GPUs (50-series).
Quality Evaluation
Below is the quality and similarity evaluated with 256 samples from MJHQ-30K dataset. (BF16 is the unqauntized model. INT W4A4 is INT4 and NVFP4 is FP4)
| Model | Precision | Method | FID | IR | LPIPS | PSNR |
|---|---|---|---|---|---|---|
| Project0 (34 Steps) | BF16 | -- | 135.52 | 0.842 | -- | -- |
| INT W4A4 | SVDQ | 135.50 | 0.795 | 0.255 | 20.64 | |
| NVFP4 | SVDQ | 134.19 | 0.849 | 0.219 | 21.30 |
| Model | Precision | Method | FID | IR | LPIPS | PSNR |
|---|---|---|---|---|---|---|
| Project0 (34 Steps) | BF16 | -- | 135.52 | 0.842 | -- | -- |
| INT W4A4 | SVDQ | 135.50 | 0.795 | 0.255 | 20.64 | |
| NVFP4 | SVDQ | 134.19 | 0.849 | 0.219 | 21.30 |
If you find these useful please consider: 