lmms_eval/tasks/olympiadbench/olympiadbench_test_cn.yaml

dataset_path: lmms-lab/OlympiadBench
dataset_kwargs:
  token: True
task : "olympiadbench_test_cn"
test_split: test_cn
output_type: generate_until
doc_to_visual: !function cn_utils.olympiadbench_doc_to_visual
doc_to_text: !function cn_utils.olympiadbench_doc_to_text
doc_to_target: "answer"
generation_kwargs:
  until:
    - "ASSISTANT:"
  max_new_tokens: 1024
  temperature: 0
  top_p: 0
  num_beams: 1
  do_sample: false
process_results: !function cn_utils.olympiadbench_process_results
metric_list:
  - metric: submission
    aggregation: !function cn_utils.olympiadbench_aggregate_results
    higher_is_better: true
  - metric: exact_match
    aggregation: mean
    higher_is_better: true