[Diffusion] [NPU] Wan2.2-T2V-A14B-Diffusers modelslim quantization support (#17996)
Co-authored-by: ronnie_zheng <zl19940307@163.com>
This commit is contained in:
co-authored by
ronnie_zheng
parent
f8d4eb7022
commit
5297b02c88
@@ -105,7 +105,7 @@ class NPUW8A8Int8DynamicLinearMethod(_NPULinearMethodBase):
|
||||
quant_out,
|
||||
layer.weight,
|
||||
layer.weight_scale,
|
||||
pertoken_scale=dynamic_scale,
|
||||
pertoken_scale=dynamic_scale.flatten(),
|
||||
bias=bias,
|
||||
output_dtype=original_dtype,
|
||||
)
|
||||
@@ -137,7 +137,7 @@ class NPU_W4A4DynamicLinearMethod(_NPULinearMethodBase):
|
||||
quant_out,
|
||||
layer.weight,
|
||||
layer.weight_scale,
|
||||
pertoken_scale=dynamic_scale,
|
||||
pertoken_scale=dynamic_scale.flatten(),
|
||||
bias=bias,
|
||||
output_dtype=original_dtype,
|
||||
)
|
||||
|
||||
Reference in New Issue
Block a user