simcop2387/AutoGPTQ
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages 1 Wiki Activity

Change referenes to 'group_size' to 'groupsize' to match rest of this file

Browse source
This commit is contained in:
TheBloke 2023-05-05 13:21:13 +01:00
parent b19c59541b
commit f64c71e779
1 changed files with 2 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
auto_gptq/modeling/_utils.py
View file

@ -82,7 +82,7 @@ def pack_model(
model, model,
quantizers, quantizers,
bits, bits,
group_size, groupsize,
use_triton=False, use_triton=False,
use_cuda_fp16=True, use_cuda_fp16=True,
desc_act=False, desc_act=False,
@ -103,7 +103,7 @@ def pack_model(
logger.info('Packing model...') logger.info('Packing model...')
layers = find_layers(model) layers = find_layers(model)
layers = {n: layers[n] for n in quantizers} layers = {n: layers[n] for n in quantizers}
make_quant(model, quantizers, bits, group_size, use_triton=use_triton, use_cuda_fp16=use_cuda_fp16, desc_act=desc_act) make_quant(model, quantizers, bits, groupsize, use_triton=use_triton, use_cuda_fp16=use_cuda_fp16, desc_act=desc_act)
qlayers = find_layers(model, [QuantLinear]) qlayers = find_layers(model, [QuantLinear])
for name in qlayers: for name in qlayers:
logger.info(name) logger.info(name)