fix bugs

intel · Jun 4, 2024 · f537e06 · f537e06
1 parent 625f5e9
commit f537e06
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 3 deletions.
diff --git a/auto_round/autoround.py b/auto_round/autoround.py
@@ -22,7 +22,7 @@
 from torch import autocast
 
 from .calib_dataset import get_dataloader
-from .quantizer import WrapperMultiblock, wrapper_block, unwrapper_block
+from .quantizer import WrapperMultiblock, wrapper_block, unwrapper_block, WrapperLinear, unwrapper_layer
 from .special_model_handler import check_hidden_state_dim, check_share_attention_mask
 from .utils import (
     CpuInfo,

diff --git a/auto_round/quantizer.py b/auto_round/quantizer.py
@@ -135,8 +135,7 @@ def quant_weight_actor(weight, num_bits, sym, v, min_scale, max_scale, scale_dty
 
 
 def quant_weight(
-        weight, num_bits=4, group_size=-1, sym=False, v=0, min_scale=1.0, max_scale=1.0, scale_dtype=torch.float16,
-        **kwargs
+        weight, num_bits=4, group_size=-1, sym=False, v=0, min_scale=1.0, max_scale=1.0, scale_dtype=torch.float16
 ):
     """Quantizes and dequantizes weight, handing the group size issue .