biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Doc] Fix typo (#11666)

Browse Source 
Signed-off-by: Kazuhiro Serizawa <nserihiro@gmail.com>
This commit is contained in:
Kazuhiro Serizawa
2025-01-01 17:10:10 +09:00
committed by GitHub
parent f962f426bc
commit 6d70198b17
2 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/layers/rejection_sampler.py
View File

@@ -39,7 +39,7 @@ class RejectionSampler(SpecDecodeStochasticBaseSampler):
strict_mode: Whether or not to perform shape/device/dtype checks
during sampling. This catches correctness issues but adds
nontrivial latency.
use_falshinfer: We will use this parameter to determine whether
use_flashinfer: We will use this parameter to determine whether
to use the FlashInfer rejection sampling kernel or not. If it's
None, we will use the default value from the environment variable.
This parameter is only used for testing purposes.