xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-04-16 20:17:08 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon 2025-09-01 02:10:48 -07:00
parent 4c2a337e67
commit b16e2d9602
1 changed files with 5 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
vllm/v1/worker/gpu_worker_states.py
View File

@ -174,13 +174,16 @@ class RequestState:
num_computed_tokens: int,
sampling_params: SamplingParams,
) -> None:
assert len(self.free_indices) > 0, "No free space in GPU worker states"
req_idx = self.free_indices.pop()
self.req_id_to_index[req_id] = req_idx
self.index_to_req_id[req_idx] = req_id
self.num_prompt_tokens.np[req_idx] = len(prompt_token_ids)
prompt_len = len(prompt_token_ids)
self.num_prompt_tokens.np[req_idx] = prompt_len
self.num_tokens.np[req_idx] = prompt_len
self.token_ids.np[req_idx, :prompt_len] = prompt_token_ids
self.num_computed_tokens.np[req_idx] = num_computed_tokens
self.append_token_ids(req_idx, prompt_token_ids)
self.temperature.np[req_idx] = sampling_params.temperature
if sampling_params.sampling_type == SamplingType.GREEDY: