anodyine/vllm-omni
1
0 Fork 0
Code Issues 4 Pull Requests 2 Actions Packages Projects Releases Wiki Activity

[NPU][Quant] Add W4A4 MXFP4 online & MXFP4 dual-scale online/offline quantization support for Wan2.2 T2V / I2V inference on Ascend NPU (#3578)

Browse Source 
Signed-off-by: hyh_hh <huyinghong1@huawei.com>
Co-authored-by: hyh_hh <huyinghong1@huawei.com>
This commit is contained in:
hxhhhlalalaandhyh_hh authored and GitHub committed 2026-05-20 06:21:29 +00:00
1 parent 109bba8cdf
commit 9dd36e3afd
14 files changed
+3646 -61
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

examples/offline_inference/image_to_video/image_to_video.py
+2 -2
View File
@@ -173,8 +173,8 @@ def parse_args() -> argparse.Namespace:
"--quantization",
type=str,
default=None,
choices=["fp8", "mxfp8", "int8", "gguf"],
help="Quantization method for the transformer. mxfp8: W8A8 MXFP8 online quant (NPU). fp8: online FP8 (GPU).",
choices=["fp8", "mxfp8", "mxfp4", "mxfp4_dualscale", "int8", "gguf"],
help="Quantization method for the transformer. mxfp8: W8A8 MXFP8 (NPU). mxfp4: W4A4 MXFP4 (NPU). mxfp4_dualscale: W4A4 MXFP4 dual-scale + BF16 fallback mixed (NPU). fp8: online FP8 (GPU).",
)
# Distributed and parallel execution
examples/offline_inference/text_to_video/text_to_video.py
+2 -2
View File
@@ -163,8 +163,8 @@ def parse_args() -> argparse.Namespace:
"--quantization",
type=str,
default=None,
choices=["fp8", "mxfp8", "int8", "gguf"],
help="Quantization method for the transformer. mxfp8: W8A8 MXFP8 online quant (NPU). fp8: online FP8 (GPU).",
choices=["fp8", "mxfp8", "mxfp4", "mxfp4_dualscale", "int8", "gguf"],
help="Quantization method for the transformer. mxfp8: W8A8 MXFP8 (NPU). mxfp4: W4A4 MXFP4 (NPU). mxfp4_dualscale: W4A4 MXFP4 dual-scale + BF16 fallback mixed (NPU). fp8: online FP8 (GPU).",
)
# Distributed and parallel execution