cleanup

gptqmodel/models/auto.py

@@ -30,6 +30,7 @@
 from .minicpm3 import MiniCPM3GPTQ
 from .mistral import MistralGPTQ
 from .mixtral import MixtralGPTQ
+from .mllama import MLlamaGPTQ
 from .moss import MOSSGPTQ
 from .mpt import MPTGPTQ
 from .opt import OPTGPTQ
@@ -43,7 +44,6 @@
 from .starcoder2 import Starcoder2GPTQ
 from .xverse import XverseGPTQ
 from .yi import YiGPTQ
-from .mllama import MLlamaGPTQ
 
 MODEL_MAP = {
     "bloom": BloomGPTQ,

tests/test_dynamic.py

@@ -9,10 +9,10 @@
 
 from datasets import load_dataset  # noqa: E402
 from gptqmodel import BACKEND, GPTQModel  # noqa: E402
-from gptqmodel.quantization import QuantizeConfig  # noqa: E402
-from gptqmodel.utils import Perplexity  # noqa: E402
 from gptqmodel.nn_modules.qlinear.qlinear_marlin_inference import MarlinInferenceQuantLinear
 from gptqmodel.nn_modules.qlinear.qlinear_tritonv2 import TritonV2QuantLinear
+from gptqmodel.quantization import QuantizeConfig  # noqa: E402
+from gptqmodel.utils import Perplexity  # noqa: E402
 from parameterized import parameterized
 from transformers import AutoTokenizer  # noqa: E402
 

tests/test_q4_marlin.py

@@ -1,6 +1,7 @@
 # -- do not touch
 import os
 import tempfile
+
 from datasets import load_dataset
 
 os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"