423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-14 10:56:56 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix #4022

Browse Source 
Former-commit-id: 24e1c0e2ee
This commit is contained in:
hiyouga
2024-06-03 18:38:36 +08:00
parent af7748139a
commit d0ceb1b091
3 changed files with 3 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/llamafactory/chat/vllm_engine.py
View File

@@ -158,12 +158,10 @@ class VllmEngine(BaseEngine):
)
result_generator = self.model.generate(
prompt=None,
inputs={"prompt_token_ids": prompt_ids, "multi_modal_data": multi_modal_data},
sampling_params=sampling_params,
request_id=request_id,
prompt_token_ids=prompt_ids,
lora_request=self.lora_request,
multi_modal_data=multi_modal_data,
)
return result_generator

2
src/llamafactory/hparams/parser.py
View File

@@ -94,7 +94,7 @@ def _check_extra_dependencies(
require_version("mixture-of-depth>=1.1.6", "To fix: pip install mixture-of-depth>=1.1.6")
if model_args.infer_backend == "vllm":
require_version("vllm>=0.4.1", "To fix: pip install vllm>=0.4.1")
require_version("vllm>=0.4.3", "To fix: pip install vllm>=0.4.3")
if finetuning_args.use_galore:
require_version("galore_torch", "To fix: pip install galore_torch")