A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 9, 07:30 UTC
[Bugfix] Fix typo in read_offset variable name (#33426)
Signed-off-by: Alberto Ferrer <albertof@barrahome.org>
Alberto Ferrer committed
Jan 30, 2026 at 19:26 UTC
64a40a7ab4d0053830fae04c83763fa67f2183e6
1 file changed
+1
-1
vllm/v1/engine/detokenizer.py
+1
-1
@@ -280,7 +280,7 @@ class SlowIncrementalDetokenizer(BaseIncrementalDetokenizer):
280
# Prompt embedding requests cannot be detokenized, in general.
281
self.tokens = [""] * self.prompt_len
282
self.prefix_offset = 0
283
- self.read_offest = 0
283
+ self.read_offset = 0
284
285
self.token_ids.extend(request.prompt_token_ids or [0] * self.prompt_len)
286