[diffusion] [AMD] feat: support online MXFP4 and fp8 quantization (#21431)

Co-authored-by: Bowen Bao <bowenbao@amd.com> Co-authored-by: HAI <hixiao@gmail.com>
2026-05-14 08:52:01 +08:00
parent af7511e0e8
commit 34c0029f0a
10 changed files with 417 additions and 17 deletions
@@ -101,6 +101,8 @@ For quantized transformer checkpoints, prefer:
 - `--model-path` for the base pipeline
 - `--transformer-path` for a quantized `transformers` transformer component folder
 - `--transformer-weights-path` for a quantized safetensors file, directory, or repo
+- `--quantization` for online quantization (apply quantization to unquantized models at load time, activations are quantized dynamically)
+- `--quantization-ignored-layers` layer name patterns to keep unquantized (e.g. `attention.to_`)

 See [Quantization](../quantization.md) for supported quantization families and examples.