[Performance] V1 Pooling Models E2E Performance Optimization (#23162)
Signed-off-by: wang.yuqi <noooop@126.com>
This commit is contained in:
@@ -713,7 +713,7 @@ class InputBatch:
|
||||
|
||||
return PoolingMetadata(
|
||||
prompt_lens=torch.from_numpy(
|
||||
self.num_prompt_tokens[:self.num_reqs]).to(self.device),
|
||||
self.num_prompt_tokens[:self.num_reqs]),
|
||||
prompt_token_ids=self.sampling_metadata.prompt_token_ids,
|
||||
pooling_params=pooling_params,
|
||||
)
|
||||
|
||||
Reference in New Issue
Block a user