This is a small python based tool, that takes in any Precision Model (like BF16, FP8, INT8) and converts it to any other Precision. It utilized ComfyuiKitchen, and takes a few seconds to minutes, depending on the hardware.

How to Use:

1. Extract the zip file.

2. Drag and drop any .safetensors model onto convert_to_<PRECISION>.bat (or convert_custom.bat).

Requirements: Python 3.10+ & NVIDIA CUDA GPU pip install torch safetensors comfy-kitchen[cublas])