api_key: EMPTY
api_url: http://127.0.0.1:8801/v1/chat/completions
chat_template: null
dataset_args:
  arc:
    dataset_id: modelscope/ai2_arc
    eval_split: test
    few_shot_num: 0
    few_shot_random: false
    name: arc
    prompt_template: ''
    subset_list:
    - ARC-Easy
    - ARC-Challenge
    train_split: train
  ceval:
    dataset_id: modelscope/ceval-exam
    eval_split: val
    few_shot_num: 0
    few_shot_random: false
    name: ceval
    prompt_template: null
    subset_list:
    - computer_network
    - operating_system
    - computer_architecture
    - college_programming
    train_split: dev
  gsm8k:
    dataset_id: modelscope/gsm8k
    eval_split: test
    few_shot_num: 4
    few_shot_random: false
    name: gsm8k
    prompt_template: ''
    subset_list:
    - main
    train_split: train
  humaneval:
    dataset_id: modelscope/humaneval
    eval_split: test
    few_shot_num: 0
    few_shot_random: false
    name: humaneval
    prompt_template: 'Complete the following python code:

      '
    subset_list:
    - openai_humaneval
    train_split: null
  ifeval:
    dataset_id: opencompass/ifeval
    eval_split: train
    few_shot_num: 0
    few_shot_random: false
    name: ifeval
    prompt_template: ''
    subset_list:
    - default
    train_split: null
dataset_dir: /root/.cache/modelscope/datasets
dataset_hub: modelscope
datasets:
- ifeval
- humaneval
- gsm8k
- arc
- ceval
debug: false
dry_run: false
eval_backend: Native
eval_config: null
eval_type: service
generation_config:
  do_sample: false
  max_length: 2048
  max_new_tokens: 512
  temperature: 1.0
  top_k: 50
  top_p: 1.0
limit: 10
mem_cache: false
model: Qwen2.5-0.5B-Instruct
model_args:
  device: auto
  precision: torch.float16
  revision: master
model_id: Qwen2.5-0.5B-Instruct
outputs: null
seed: 42
stage: all
template_type: null
use_cache: null
work_dir: ./outputs/20250117_154119