Skip to content

remove ark support for cpu #1206

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
wenhuach21 merged 2 commits into from
Dec 30, 2025
Merged

remove ark support for cpu #1206

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
76f5340
update
wenhuach21 Dec 30, 2025
121b999
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Dec 30, 2025
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
121 changes: 76 additions & 45 deletions auto_round/inference/backend.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -439,21 +439,21 @@ def fp8_static_scheme_checker(
requirements=["autoawq", "transformers"],
)

BackendInfos["auto_round_kernel"] = BackendInfo(
device=["cpu"],
sym=[True, False],
packing_format=GPTQ_FORMAT_NO_ZP,
bits=[2, 4, 8],
group_size=None,
priority=6,
checkers=[ark_feature_checker],
alias=["ark"],
compute_dtype=["float32", "float16"],
data_type=["int"],
act_bits=WOQ_DEFAULT_ACT_BITS,
requirements=["torch>=2.9.0", "auto_round_kernel"],
systems=["linux"],
)
# BackendInfos["auto_round_kernel"] = BackendInfo(
# device=["cpu"],
# sym=[True, False],
# packing_format=GPTQ_FORMAT_NO_ZP,
# bits=[2, 4, 8],
# group_size=None,
# priority=6,
# checkers=[ark_feature_checker],
# alias=["ark"],
# compute_dtype=["float32", "float16"],
# data_type=["int"],
# act_bits=WOQ_DEFAULT_ACT_BITS,
# requirements=["torch>=2.9.0", "auto_round_kernel"],
# systems=["linux"],
# )

BackendInfos["auto_round_kernel_xpu"] = BackendInfo(
device=["xpu"],
Expand All @@ -471,11 +471,27 @@ def fp8_static_scheme_checker(
systems=["linux"],
)

BackendInfos["auto_round_kernel_zp"] = BackendInfo(
device=["cpu"],
sym=[True, False],
# BackendInfos["auto_round_kernel_zp"] = BackendInfo(
# device=["cpu"],
# sym=[True, False],
# packing_format=GPTQ_FORMAT,
# bits=[2, 4, 8],
# group_size=None,
# priority=6,
# checkers=[ark_feature_checker],
# alias=["ark"],
# compute_dtype=["float32", "float16"],
# data_type=["int"],
# act_bits=WOQ_DEFAULT_ACT_BITS,
# requirements=["torch>=2.9.0", "auto_round_kernel"],
# systems=["linux"],
# )

BackendInfos["auto_round_kernel_zp_xpu"] = BackendInfo(
device=["xpu"],
sym=[True],
packing_format=GPTQ_FORMAT,
bits=[2, 4, 8],
bits=[4, 8],
group_size=None,
priority=6,
checkers=[ark_feature_checker],
Expand All @@ -487,10 +503,26 @@ def fp8_static_scheme_checker(
systems=["linux"],
)

BackendInfos["auto_round_kernel_zp_xpu"] = BackendInfo(
# BackendInfos["auto_round_kernel_awq"] = BackendInfo(
# device=["cpu"],
# sym=[True, False],
# packing_format=AWQ_FORMAT,
# bits=[2, 4, 8],
# group_size=None,
# priority=6,
# checkers=[ark_feature_checker],
# alias=["ark"],
# compute_dtype=["float32", "float16"],
# data_type=["int"],
# act_bits=WOQ_DEFAULT_ACT_BITS,
# requirements=["torch>=2.9.0", "auto_round_kernel"],
# systems=["linux"],
# )

BackendInfos["auto_round_kernel_awq_xpu"] = BackendInfo(
device=["xpu"],
sym=[True],
packing_format=GPTQ_FORMAT,
packing_format=AWQ_FORMAT,
bits=[4, 8],
group_size=None,
priority=6,
Expand All @@ -503,42 +535,40 @@ def fp8_static_scheme_checker(
systems=["linux"],
)

BackendInfos["auto_round_kernel_awq"] = BackendInfo(
BackendInfos["ipex_gptq_cpu"] = BackendInfo(
device=["cpu"],
sym=[True, False],
packing_format=AWQ_FORMAT,
bits=[2, 4, 8],
packing_format=GPTQ_FORMAT,
bits=[4],
group_size=None,
priority=6,
checkers=[ark_feature_checker],
alias=["ark"],
compute_dtype=["float32", "float16"],
priority=5,
checkers=[],
compute_dtype=["float16", "bfloat16"],
data_type=["int"],
act_bits=WOQ_DEFAULT_ACT_BITS,
requirements=["torch>=2.9.0", "auto_round_kernel"],
systems=["linux"],
alias=["ipex"],
requirements=["torch<2.9", "intel-extension-for-pytorch>=2.5"],
)

BackendInfos["auto_round_kernel_awq_xpu"] = BackendInfo(
BackendInfos["ipex_gptq"] = BackendInfo(
device=["xpu"],
sym=[True],
packing_format=AWQ_FORMAT,
bits=[4, 8],
sym=[True, False],
packing_format=GPTQ_FORMAT,
bits=[4],
group_size=None,
priority=6,
checkers=[ark_feature_checker],
alias=["ark"],
compute_dtype=["float32", "float16"],
priority=5,
checkers=[],
compute_dtype=["float16", "bfloat16"],
data_type=["int"],
act_bits=WOQ_DEFAULT_ACT_BITS,
requirements=["torch>=2.9.0", "auto_round_kernel"],
systems=["linux"],
alias=["ipex"],
requirements=["intel-extension-for-pytorch>=2.5"],
)

BackendInfos["ipex_gptq"] = BackendInfo(
device=["cpu", "xpu"],
BackendInfos["ipex_awq_cpu"] = BackendInfo(
device=["cpu"],
sym=[True, False],
packing_format=GPTQ_FORMAT,
packing_format=AWQ_FORMAT,
bits=[4],
group_size=None,
priority=5,
Expand All @@ -547,11 +577,12 @@ def fp8_static_scheme_checker(
data_type=["int"],
act_bits=WOQ_DEFAULT_ACT_BITS,
alias=["ipex"],
requirements=["intel-extension-for-pytorch>=2.5"],
requirements=["torch<2.9", "intel-extension-for-pytorch>=2.5"],
)


BackendInfos["ipex_awq"] = BackendInfo(
device=["cpu", "xpu"],
device=["cpu"],
sym=[True, False],
packing_format=AWQ_FORMAT,
bits=[4],
Expand Down