Skip to content

GitLab

Unverified Commit d79ced32 authored by Wen Sun's avatar Wen Sun Committed by GitHub
Browse files

Fix 'Actor methods cannot be called directly' when using `--engine-use-ray` (#2664) 

* fix: engine-useray complain

* fix: typo
parent ab406446
Hide whitespace changes
Inline Side-by-side
Showing with 13 additions and 5 deletions
vllm/engine/async_llm_engine.py
View file @ d79ced32
...@@ -447,11 +447,19 @@ class AsyncLLMEngine: ...@@ -447,11 +447,19 @@ class AsyncLLMEngine:
if arrival_time is None: if arrival_time is None:
arrival_time = time.time() arrival_time = time.time()
prompt_token_ids = await self.engine.encode_request_async(
request_id=request_id, if self.engine_use_ray:
prompt=prompt, prompt_token_ids = await self.engine.encode_request_async.remote(
prompt_token_ids=prompt_token_ids, request_id=request_id,
lora_request=lora_request) prompt=prompt,
prompt_token_ids=prompt_token_ids,
lora_request=lora_request)
else:
prompt_token_ids = await self.engine.encode_request_async(
request_id=request_id,
prompt=prompt,
prompt_token_ids=prompt_token_ids,
lora_request=lora_request)
stream = self._request_tracker.add_request( stream = self._request_tracker.add_request(
request_id, request_id,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment