From a989ca2bf6398b63607637892240375c63d76a9e Mon Sep 17 00:00:00 2001 From: Shanshan Shen <467638484@qq.com> Date: Fri, 21 Mar 2025 16:58:28 +0800 Subject: [PATCH] [Bugfix] Add int8 torch dtype for KVCache (#15260) Signed-off-by: shen-shanshan <467638484@qq.com> --- vllm/utils.py | 1 + 1 file changed, 1 insertion(+) diff --git a/vllm/utils.py b/vllm/utils.py index b723637b25013..55ee044b482cd 100644 --- a/vllm/utils.py +++ b/vllm/utils.py @@ -153,6 +153,7 @@ STR_DTYPE_TO_TORCH_DTYPE = { "fp8": torch.uint8, "fp8_e4m3": torch.uint8, "fp8_e5m2": torch.uint8, + "int8": torch.int8, } TORCH_DTYPE_TO_NUMPY_DTYPE = {