[Misc] Reorganize inputs (#35182)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2026-03-25 10:22:54 -07:00
committed by GitHub
parent 678b3c99e8
commit ba2f0acc2d
142 changed files with 1212 additions and 1342 deletions
+9 -9
View File
@@ -20,7 +20,7 @@ from vllm.entrypoints.serve.tokenize.protocol import (
TokenizeResponse,
TokenizerInfoResponse,
)
from vllm.inputs import TokensPrompt, token_inputs
from vllm.inputs import TokensPrompt, tokens_input
from vllm.logger import init_logger
from vllm.tokenizers import TokenizerLike
@@ -79,7 +79,7 @@ class OpenAIServingTokenization(OpenAIServing):
if error_check_ret is not None:
return error_check_ret
_, engine_prompts = await self.openai_serving_render.preprocess_chat(
_, engine_inputs = await self.openai_serving_render.preprocess_chat(
request,
request.messages,
default_template=self.chat_template,
@@ -88,22 +88,22 @@ class OpenAIServingTokenization(OpenAIServing):
tool_dicts=tool_dicts,
)
else:
engine_prompts = await self.openai_serving_render.preprocess_completion(
engine_inputs = await self.openai_serving_render.preprocess_completion(
request,
prompt_input=request.prompt,
prompt_embeds=None,
)
input_ids: list[int] = []
for engine_prompt in engine_prompts:
for engine_input in engine_inputs:
self._log_inputs(
request_id,
engine_prompt,
engine_input,
params=None,
lora_request=lora_request,
)
prompt_components = self._extract_prompt_components(engine_prompt)
prompt_components = self._extract_prompt_components(engine_input)
if prompt_components.token_ids is not None:
input_ids.extend(prompt_components.token_ids)
@@ -134,16 +134,16 @@ class OpenAIServingTokenization(OpenAIServing):
self._log_inputs(
request_id,
token_inputs(request.tokens),
tokens_input(request.tokens),
params=None,
lora_request=lora_request,
)
engine_prompt = await self.renderer.tokenize_prompt_async(
tok_prompt = await self.renderer.tokenize_prompt_async(
TokensPrompt(prompt_token_ids=request.tokens),
request.build_tok_params(self.model_config),
)
prompt_text = engine_prompt["prompt"] # type: ignore[typeddict-item]
prompt_text = tok_prompt["prompt"] # type: ignore[typeddict-item]
return DetokenizeResponse(prompt=prompt_text)