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
remove useless statements (#1408)
Wang Ran (汪然) committed
Oct 20, 2023 at 23:52 UTC
d189170b6c5a143e493c3f5cb7e8c976e8da62c7
1 file changed
-1
vllm/model_executor/input_metadata.py
-1
@@ -66,7 +66,6 @@ class InputMetadata:
66
else:
67
self.max_num_blocks_per_seq = 0
68
assert block_tables.shape[0] == self.num_generation_tokens
69
- assert context_lens.shape[0] == self.num_generation_tokens
69
70
# Set during the execution of the first attention op.
71
self.attn_bias: Optional[AttentionBias] = None