A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 14:45 UTC
[Misc] Delete redundant code (#16530)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com> Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Jee Jee Li committed
Apr 12, 2025 at 19:21 UTC
3cdc57669f231a2c2bdbebb0480164d10cf89edb
2 files changed
-14
examples/offline_inference/vision_language.py
-7
@@ -1096,13 +1096,6 @@ def main(args):
1096
}
1097
llm = LLM(**engine_args)
1098
1099
- # To maintain code compatibility in this script, we add LoRA here.
1100
- # You can also add LoRA using:
1101
- # llm.generate(prompts, lora_request=lora_request,...)
1102
- if req_data.lora_requests:
1103
- for lora_request in req_data.lora_requests:
1104
- llm.llm_engine.add_lora(lora_request=lora_request)
1105
-
1099
# Don't want to check the flag multiple times, so just hijack `prompts`.
1100
prompts = req_data.prompts if args.use_different_prompt_per_request else [
1101
req_data.prompts[0]
examples/offline_inference/vision_language_multi_image.py
-7
@@ -695,13 +695,6 @@ def run_chat(model: str, question: str, image_urls: list[str],
695
engine_args = asdict(req_data.engine_args) | {"seed": seed}
696
llm = LLM(**engine_args)
697
698
- # To maintain code compatibility in this script, we add LoRA here.
699
- # You can also add LoRA using:
700
- # llm.generate(prompts, lora_request=lora_request,...)
701
- if req_data.lora_requests:
702
- for lora_request in req_data.lora_requests:
703
- llm.llm_engine.add_lora(lora_request=lora_request)
704
-
698
sampling_params = SamplingParams(temperature=0.0,
699
max_tokens=256,
700
stop_token_ids=req_data.stop_token_ids)