[Bugfix] Fix gptq failure on T4s (#7264)

2024-08-07 16:05:37 -04:00
parent 469b3bc538
commit 311f743831
3 changed files with 14 additions and 15 deletions
--- a/vllm/model_executor/layers/quantization/awq_marlin.py
+++ b/vllm/model_executor/layers/quantization/awq_marlin.py
@@ -126,8 +126,7 @@ class AWQMarlinConfig(QuantizationConfig):

        return check_marlin_supported(quant_type=cls.TYPE_MAP[num_bits],
                                      group_size=group_size,
-                                      has_zp=has_zp,
-                                      min_capability=cls.get_min_capability())
+                                      has_zp=has_zp)


 class AWQMarlinLinearMethod(LinearMethodBase):