huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-11 12:24:53 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix incorrect use of bias in awq

Browse Source
This commit is contained in:
chenxichen 2023-12-27 03:25:47 +00:00
parent dd304cf14c
commit 74d9dfa89e
2 changed files with 2 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
server/text_generation_server/utils/awq/quantize/qmodule.py
View File

@ -35,10 +35,7 @@ class WQLinear(nn.Module):
self.qweight = qweight self.qweight = qweight
self.qzeros = qzeros self.qzeros = qzeros
self.scales = scales self.scales = scales
if bias:
self.bias = bias self.bias = bias
else:
self.bias = None
@torch.no_grad() @torch.no_grad()
def forward(self, x): def forward(self, x):

2
server/text_generation_server/utils/layers.py
View File

@ -335,7 +335,7 @@ def get_linear(weight, bias, quantize):
qweight=qweight, qweight=qweight,
qzeros=qzeros, qzeros=qzeros,
scales=scales, scales=scales,
bias=bias is not None, bias=bias,
) )
else: else:
raise NotImplementedError(f"Quantization `{quantize}` is not implemented yet.") raise NotImplementedError(f"Quantization `{quantize}` is not implemented yet.")