xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-07-23 20:37:24 +08:00
Code Issues Packages Projects Releases Wiki Activity

flash_infer missing out dtype bug fix

Browse Source 
Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>
This commit is contained in:
vllmellm 2025-11-03 08:02:45 +00:00
parent 7794009661
commit a8010c7b1c
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/layers/quantization/kernels/scaled_mm/utils.py
View File

@ -17,7 +17,7 @@ def apply_weights_fp8(
x_s: torch.Tensor, x_s: torch.Tensor,
bias: torch.Tensor, bias: torch.Tensor,
x_s_ub: torch.Tensor | None, x_s_ub: torch.Tensor | None,
maybe_out_dtype: torch.dtype | None, maybe_out_dtype: torch.dtype | None = None,
) -> torch.Tensor: ) -> torch.Tensor:
# ops.scaled_fp8_quant supports both dynamic and static quant. # ops.scaled_fp8_quant supports both dynamic and static quant.
# If dynamic, layer.input_scale is None and x_s computed from x. # If dynamic, layer.input_scale is None and x_s computed from x.