@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 2e8e49fce3
Fork 0

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 17, 06:45 UTC

[Fix] Remove false assertion (#1222)

Woosuk Kwon committed Sep 28, 2023 at 10:52 UTC 2e8e49fce3775e7704d413b2f02da6d7c99525c9
1 file changed -2
vllm/worker/worker.py
-2
@@ -155,8 +155,6 @@ class Worker:
155 self,
156 seq_group_metadata_list: List[SequenceGroupMetadata],
157 ) -> Tuple[torch.Tensor, torch.Tensor, InputMetadata]:
158 - assert self.block_size is not None
159 -
158 seq_groups: List[Tuple[List[int], SamplingParams]] = []
159 input_tokens: List[int] = []
160 input_positions: List[int] = []