(self, model_path)
| 24 | @ray.remote(num_gpus=1) |
| 25 | class LLMModelActor: |
| 26 | def __init__(self, model_path): |
| 27 | self.model = LocalModelLLMServing_vllm( |
| 28 | hf_model_name_or_path=model_path, |
| 29 | vllm_tensor_parallel_size=1, |
| 30 | vllm_temperature=0.7, |
| 31 | vllm_top_p=0.9, |
| 32 | vllm_max_tokens=4096, |
| 33 | vllm_gpu_memory_utilization=0.85 |
| 34 | ) |
| 35 | |
| 36 | def generate_from_input(self, *args, **kwargs): |
| 37 | return self.model.generate_from_input(*args, **kwargs) |
nothing calls this directly
no test coverage detected