--- pipeline_tag: text-to-image tags: - Flux1 - SVDQaunt - SVDQ - Quantization - jib - j1b - Jib Mix Flux language: - en base_model_relation: quantized library_name: diffusers --- This repository contains Nunchaku-quantized (SVDQ) versions of [Jib Mix Flux](https://civitai.com/models/686814/jib-mix-flux), a text-to-image model based on Flux.1-Dev by [J1B](https://civitai.com/user/J1B) If you like the model, please consider liking, reviewing and tipping the creator. Model Files - [svdq-int4_r32-Jib-Mix-Flux-V12.safetensors](https://huggingface.co/spooknik/Jib-Mix-Flux-SVDQ/blob/main/svdq-int4_r32-Jib-Mix-Flux-V12.safetensors): SVDQuant INT4 (rank 32) Jib Mix Flux V12 model. For users with non-Blackwell GPUs (pre-50-series). - [svdq-fp4_r32-Jib-Mix-Flux-V12.safetensors](https://huggingface.co/spooknik/Jib-Mix-Flux-SVDQ/blob/main/svdq-fp4_r32-Jib-Mix-Flux-V12.safetensors): SVDQuant NVFP4 (rank 32) Jib Mix Flux V12 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 | |----------------------------|-----------|-----------|--------------------|-----------------|----------------------|-------------------| | Jib Flux Mix V12 (25 steps)| BF16 | -- | 170.37 | 0.946 | -- | -- | | | INT W4A4 | SVDQ | 171.64 | 0.907 | 0.253 | 20.31 | | | NVFP4 | SVDQ | 170.99 | 0.955 | 0.220 | 20.71 | If you find these useful please consider: Buy Me a Coffee at ko-fi.com