From 7d26e698978e696b48be6e88dedd3f42f0c92582 Mon Sep 17 00:00:00 2001 From: kewang-xlnx Date: Wed, 18 Dec 2024 21:06:14 -0700 Subject: [PATCH] delete comments of get_cache_scale in compressed_tensors.py Signed-off-by: kewang-xlnx --- .../quantization/compressed_tensors/compressed_tensors.py | 4 ---- 1 file changed, 4 deletions(-) diff --git a/vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors.py b/vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors.py index b87e102f9795a..fc35522702d4d 100644 --- a/vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors.py +++ b/vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors.py @@ -412,10 +412,6 @@ def get_scheme( self._check_scheme_supported(scheme.get_min_capability()) return scheme - # move the get_compressed_tensors_cache_scale method from - # utils.py to instance method of CompressedTensorsConfig - # class. By doing this, different QuantizationConfig - # classes can implement their own get_cache_scale method. def get_cache_scale(self, name: str) -> Optional[str]: """ Check whether the param name matches the format for k/v cache scales