Skip to content

fix: Update QuantLinear import to GPTQLoraLinear #146

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
willmj merged 4 commits into from
Jun 10, 2025
Merged

fix: Update QuantLinear import to GPTQLoraLinear #146

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
8947e96
fix: Update QuantLinear import to GPTQLoraLinear
Luka-D Jun 5, 2025
accf4df
fix: Specified peft version
Luka-D Jun 6, 2025
55a9e08
peft lower limit
willmj Jun 10, 2025
19003e7
fix: remove deps from plugins
willmj Jun 10, 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
4 changes: 2 additions & 2 deletions plugins/accelerated-peft/src/fms_acceleration_peft/autogptq_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@
ModelPatcherTrigger,
)
from peft import LoraConfig
from peft.tuners.lora.gptq import QuantLinear as LoraLinearGPTQ
from peft.tuners.lora.gptq import GPTQLoraLinear
import torch

# these parameters are to be patched for triton v2
Expand Down Expand Up @@ -162,7 +162,7 @@ def create_new_module_peft(
# to be installed
new_module = None
if isinstance(target, target_cls):
new_module = LoraLinearGPTQ(
new_module = GPTQLoraLinear(
target, adapter_name, lora_config=lora_config, **kwargs
)

Expand Down
4 changes: 2 additions & 2 deletions plugins/accelerated-peft/src/fms_acceleration_peft/gptqmodel/utils/peft.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@
from peft.mapping import PEFT_TYPE_TO_CONFIG_MAPPING
from peft.peft_model import PEFT_TYPE_TO_MODEL_MAPPING
from peft.tuners.lora import LoraConfig, LoraModel
from peft.tuners.lora.gptq import QuantLinear as LoraLinearGPTQ
from peft.tuners.lora.gptq import GPTQLoraLinear
import torch

# Local
Expand Down Expand Up @@ -68,7 +68,7 @@ def _create_new_module(
# to be installed
new_module = None
if isinstance(target, target_cls):
new_module = LoraLinearGPTQ(
new_module = GPTQLoraLinear(
target, adapter_name, lora_config=lora_config, **kwargs
)

Expand Down
2 changes: 1 addition & 1 deletion plugins/framework/pyproject.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@ classifiers=[
dependencies = [
"numpy<2.0", # numpy needs to be bounded due to incompatiblity with current torch<2.3
"torch>2.2",
"peft<=0.14.0", # QuantLinear is not available for peft version > 0.14.0
"peft>=0.15.0",
"accelerate",
"pandas",
]
Expand Down
2 changes: 1 addition & 1 deletion plugins/fused-ops-and-kernels/tests/test_fused_ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,7 @@
if _is_package_available("auto_gptq"):
# pylint: disable=ungrouped-imports
# Third Party
from peft.tuners.lora.gptq import QuantLinear as LoraGPTQLinear4bit
from peft.tuners.lora.gptq import GPTQLoraLinear as LoraGPTQLinear4bit

LORA_QUANTIZED_CLASSES[GPTQ] = LoraGPTQLinear4bit

Expand Down