Skip to content

Support woq Autotune #1921

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Jul 16, 2024
Merged

Support woq Autotune #1921

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
b920d77
support woq tuning
Kaihui-intel Jul 15, 2024
2ba9e1d
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Jul 15, 2024
65cde92
update ut
Kaihui-intel Jul 15, 2024
8d7dab5
improve coverage
Kaihui-intel Jul 16, 2024
26d6c8e
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Jul 16, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion neural_compressor/torch/algorithms/weight_only/utility.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1105,7 +1105,11 @@ def __iter__(self):
if not args:
yield kwargs
elif not kwargs:
yield args
# case: tensor
if len(args) == 1:
yield args[0]
else:
yield args
else:
yield args, kwargs

Expand Down
7 changes: 3 additions & 4 deletions neural_compressor/torch/quantization/config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -723,7 +723,7 @@ def __init__(
minmax_lr: float = None,
low_gpu_mem_usage: bool = True,
iters: int = 200,
seqlen: int = 2048,
seqlen: int = 512,
n_samples: int = 512,
sampler: str = "rand",
seed: int = 42,
Expand Down Expand Up @@ -1490,8 +1490,7 @@ def get_woq_tuning_config() -> list:
the list of WOQ quant config.
"""
RTN_G32ASYM = RTNConfig(use_sym=False, group_size=32)
AUTO_ROUND_CONFIG = AutoRoundConfig(use_sym=False, group_size=32)
GPTQ_G32ASYM = GPTQConfig(use_sym=False, group_size=32)
GPTQ_G32ASYM_DISABLE_LAST_LINEAR = GPTQConfig(use_sym=False).set_local("*.lm_head", GPTQConfig(dtype="fp32"))
GPTQ_G128ASYM = GPTQConfig(group_size=128, use_sym=False)
AWQ_G32ASYM = AWQConfig(use_sym=False, group_size=32)
return [RTN_G32ASYM, GPTQ_G32ASYM, GPTQ_G32ASYM_DISABLE_LAST_LINEAR, GPTQ_G128ASYM, AWQ_G32ASYM]
return [RTN_G32ASYM, AUTO_ROUND_CONFIG, GPTQ_G32ASYM, AWQ_G32ASYM]