[diffusion] use fa3 by default when device supports it (#783)

Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>

pyproject.toml

@@ -39,6 +39,7 @@ dependencies = [
     "cache-dit==1.2.0",
     "tqdm>=4.66.0",
     "torchsde>=0.2.6",  # Required for Stable Audio scheduler
+    "fa3-fwd", # flash attention 3, maintained by @ZJY0516
     "openai-whisper>=20250625",
     "imageio[ffmpeg]>=2.37.2",
     # "vllm==0.14.0",  # TODO: fix the entrypoints overwrite problem