Update to new version of base image

vllm/model_executor/layers/linear.py

@@ -132,7 +132,7 @@ class UnquantizedLinearMethod(LinearMethodBase):
               layer: torch.nn.Module,
               x: torch.Tensor,
               bias: Optional[torch.Tensor] = None) -> torch.Tensor:
-        if (x.shape[0] == 16384 or x.shape[0] == 15360):
+        if (x.shape[0] == 8192 or x.shape[0] == 16384 or x.shape[0] == 15360):
             if bias is None:
                 return x @ layer.weight.T
             else:

vllm/model_executor/layers/quantization/compressed_tensors/utils.py

@@ -1,4 +1,4 @@
-import re
+import regex as re
 from enum import Enum
 from typing import Any, Dict, Iterable, Optional, Union