EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

fix typo in python/sglang/srt/layers/quantization/fp8.py (#2655)

Browse Source
This commit is contained in:
HAI
2024-12-29 23:45:02 -08:00
committed by GitHub
parent 3231817861
commit e6f523b5f2
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
python/sglang/srt/layers/quantization/fp8.py
View File

@@ -280,9 +280,9 @@ class Fp8LinearMethod(LinearMethodBase):
weight_scale=layer.weight_scale_inv, weight_scale=layer.weight_scale_inv,
input_scale=None, input_scale=None,
) )
layer.weight = torch.nn.Parameter(weight, require_grad=False) layer.weight = torch.nn.Parameter(weight, requires_grad=False)
layer.weight_scale_inv = torch.nn.Parameter( layer.weight_scale_inv = torch.nn.Parameter(
weight_scale, require_grad=False weight_scale, requires_grad=False
) )
layer.input_scale = None layer.input_scale = None
return return