Kandinsky 6.0 Quantized - Int8 Convrot
Collection
5 items • Updated • 1
Original model repository: https://huggingface.co/kandinskylab/Kandinsky-6.0-Pro-5s-Diffusers.
Quantized using the comfy-quants fork for Kandinsky-6.0 https://github.com/Comfy-Org/comfy-quants/compare/main...kabachuha:comfy-quants:kand6.
To be used with https://github.com/Comfy-Org/ComfyUI/pull/16825, in native ComfyUI.
It may NOT work with the custom nodes pack provided by the original model authors.
This is a test of both the ComfyUI pull-request and of the quantization code, and to serve as a hint to the model authors.
Conversion command:
comfy-quants export-model-int8-tensorwise --config configs/kandinsky6_int8_tensorwise.yaml --source /media/kabachuha/maxx/K6/Kandinsky-6.0-Pro-5s.safetensors --out /media/kabachuha/maxx/K6/Kandinsky-6.0-Pro-5s-int8-convrot.safetensors
- Downloads last month
- 58
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for kabachuha/Kandinsky-6.0-Pro-5s-int8-convrot
Base model
kandinskylab/Kandinsky-6.0-Pro-5s-Diffusers