biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Misc] Clean up input processing (#17582)

Browse Source 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2025-05-02 23:11:53 +08:00
committed by GitHub
parent 3a500cd0b6
commit cb234955df
9 changed files with 359 additions and 285 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/entrypoints/llm.py
View File

@@ -27,7 +27,7 @@ from vllm.entrypoints.score_utils import (_cosine_similarity,
_validate_score_input_lens)
from vllm.entrypoints.utils import _validate_truncation_size
from vllm.inputs import PromptType, SingletonPrompt, TextPrompt, TokensPrompt
from vllm.inputs.parse import is_token_prompt, parse_and_batch_prompt
from vllm.inputs.parse import parse_and_batch_prompt
from vllm.logger import init_logger
from vllm.lora.request import LoRARequest
from vllm.model_executor.guided_decoding.guided_fields import (
@@ -567,10 +567,12 @@ class LLM:
mm_kwargs["mm_processor_kwargs"] = prompt[
"mm_processor_kwargs"]
if is_token_prompt(prompt):
if "prompt_token_ids" in prompt:
prompt = cast(TokensPrompt, prompt) # Needed for mypy
prompt_tokens = prompt["prompt_token_ids"]
else:
prompt_tokens = tokenizer.encode(prompt["prompt"])
instances.append(
BeamSearchInstance(prompt_tokens, logprobs=None, **mm_kwargs))