quantization_info.json

{
  "quantization": "int8",
  "method": "torchao",
  "original_model": "black-forest-labs/FLUX.1-schnell",
  "conversion_date": "2025-08-09T09:57:45.366707",
  "requires_bitsandbytes": false,
  "device": "cuda",
  "serialization_format": "pytorch",
  "notes": "TorchAO quantized models saved in PyTorch format due to storage compatibility"
}