dataset_path: lmms-lab-eval/Spatial457
test_split: test
output_type: generate_until
doc_to_visual: !function utils.spatial457_doc_to_visual
doc_to_text: !function utils.spatial457_doc_to_text
doc_to_target: "answer"

process_results: !function utils.spatial457_process_results

metric_list:
  - metric: spatial457_accuracy
    aggregation: !function utils.spatial457_aggregate_results
    higher_is_better: true

generation_kwargs:
  max_new_tokens: 512
  temperature: 0.0
  do_sample: false

lmms_eval_specific_kwargs:
  default:
    pre_prompt: ""
    post_prompt: ""

dataset_kwargs:
  token: True

metadata:
  version: 1.0
