From e1464c3a0861384974dd6cfa35f2d6ff729ab29c Mon Sep 17 00:00:00 2001 From: Isotr0py Date: Sun, 30 Nov 2025 14:04:28 +0800 Subject: [PATCH] [Quantization] Enable compressed-tensors AWQ for Turing GPU (#29732) Signed-off-by: Isotr0py --- .../compressed_tensors/schemes/compressed_tensors_wNa16.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/vllm/model_executor/layers/quantization/compressed_tensors/schemes/compressed_tensors_wNa16.py b/vllm/model_executor/layers/quantization/compressed_tensors/schemes/compressed_tensors_wNa16.py index 7f4dad70287bd..3f1b4d883b79c 100644 --- a/vllm/model_executor/layers/quantization/compressed_tensors/schemes/compressed_tensors_wNa16.py +++ b/vllm/model_executor/layers/quantization/compressed_tensors/schemes/compressed_tensors_wNa16.py @@ -79,8 +79,8 @@ class CompressedTensorsWNA16(CompressedTensorsScheme): @classmethod def get_min_capability(cls) -> int: - # ampere and up - return 80 + # Turing and up + return 75 def create_weights( self,