xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-05-23 10:44:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

[V1] Chore: cruft removal (#11724)

Browse Source
This commit is contained in:
Robert Shaw 2025-01-03 18:25:02 -05:00 committed by GitHub
parent bf0d97d786
commit ad0d567e1c
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 0 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/entrypoints/llm.py
View File

@ -225,8 +225,6 @@ class LLM:
# Logic to switch between engines is done at runtime instead of import # Logic to switch between engines is done at runtime instead of import
# to avoid import order issues # to avoid import order issues
self.engine_class = self.get_engine_class() self.engine_class = self.get_engine_class()
# TODO(rob): enable mp by default (issue with fork vs spawn)
self.llm_engine = self.engine_class.from_engine_args( self.llm_engine = self.engine_class.from_engine_args(
engine_args, usage_context=UsageContext.LLM_CLASS) engine_args, usage_context=UsageContext.LLM_CLASS)

2
vllm/v1/engine/core_client.py
View File

@ -94,8 +94,6 @@ class InprocClient(EngineCoreClient):
* pushes EngineCoreRequest directly into the EngineCore * pushes EngineCoreRequest directly into the EngineCore
* pulls EngineCoreOutputs by stepping the EngineCore * pulls EngineCoreOutputs by stepping the EngineCore
TODO: support asyncio-mode for debugging.
""" """
def __init__(self, *args, **kwargs): def __init__(self, *args, **kwargs):

4
vllm/v1/engine/llm_engine.py
View File

@ -42,8 +42,6 @@ class LLMEngine:
use_cached_outputs: bool = False, use_cached_outputs: bool = False,
multiprocess_mode: bool = False, multiprocess_mode: bool = False,
) -> None: ) -> None:
# TODO: Can we avoid this?
self.model_config = vllm_config.model_config self.model_config = vllm_config.model_config
# Tokenizer (+ ensure liveness if running in another process). # Tokenizer (+ ensure liveness if running in another process).
@ -179,8 +177,6 @@ class LLMEngine:
return request_outputs return request_outputs
# TODO(rob): Can we get rid of these?
def get_model_config(self): def get_model_config(self):
return self.model_config return self.model_config

3
vllm/v1/engine/processor.py
View File

@ -49,9 +49,6 @@ class Processor:
cache_config.enable_prefix_caching cache_config.enable_prefix_caching
self.mm_hasher = MMHasher() self.mm_hasher = MMHasher()
# TODO: run in an ThreadpoolExecutor or BackgroundProcess.
# This ideally should releases the GIL, so we should not block the
# asyncio loop while this is running.
def process_inputs( def process_inputs(
self, self,
request_id: str, request_id: str,