eval_llama_series_instruct_vllm.py

from mmengine.config import read_base

with read_base():
    from ..datasets.ARC_c.ARC_c_gen_1e0de5 import ARC_c_datasets 
    from ..datasets.ARC_e.ARC_e_gen_1e0de5 import ARC_e_datasets
    # from ..datasets.gpqa.gpqa_openai_simple_evals_gen_5aeece import gpqa_datasets
    # from ..datasets.math.math_0shot_gen_393424 import math_datasets
    from ..summarizers.example import summarizer

datasets = sum([v for k, v in locals().items() if k.endswith("_datasets") or k == 'datasets'], [])
work_dir = './outputs/llama-series/'

from opencompass.models import VLLMwithChatTemplate

settings = [
    ('llama-3.1-8b-instruct-vllm', 'meta-llama/Meta-Llama-3.1-8B-Instruct', 1),
    ('llama-3.1-70b-instruct-vllm', 'meta-llama/Meta-Llama-3.1-70B-Instruct', 4),
]

models = []
for abbr, path, num_gpus in settings:
    models.append(
        dict(
            type=VLLMwithChatTemplate,
            abbr=abbr,
            path=path,
            model_kwargs=dict(tensor_parallel_size=num_gpus),
            max_out_len=100,
            max_seq_len=2048,
            batch_size=32,
            generation_kwargs=dict(temperature=0),
            run_cfg=dict(num_gpus=num_gpus, num_procs=1),
        )
    )