I've made this model using the <a target="_blank" rel="ugc" href="https://github.com/lum3on/ComfyUI-ModelQuantizer">https://github.com/lum3on/ComfyUI-ModelQuantizer</a> nodes and the full V48 version, so it could run better in my RX6800.I know that it is also needed if you want to run TorchCompile in RTX 3000 series, so here you have it.

Chroma fp8 e5m2

Conversion directly from <a target="_blank" rel="ugc" href="https://civitai.com/models/1966367">https://civitai.com/models/1966367</a>

chromaFp8E5m2_2kQC.safetensors

2K-QC

Files with SHA256: ff02f2eb25d0939895ba48688fed503438d021c0320ea785e9f8a9f6223aed75