dataset_path: neulab/VisualPuzzles
dataset_kwargs:
  token: True
task: "VisualPuzzles_direct"
test_split: train
output_type: generate_until
doc_to_visual: !function utils.VisualPuzzles_doc_to_visual
doc_to_text: !function utils.VisualPuzzles_doc_to_text
doc_to_target: "answer"
generation_kwargs:
  max_new_tokens: 4096
  temperature: 0
  top_p: 1.0
  num_beams: 1
  do_sample: false
metric_list:
  - metric: exact_match
    aggregation: mean
    higher_is_better: true
    ignore_case: true
    ignore_punctuation: true
process_results: !function utils.VisualPuzzles_process_result
metadata:
  - version: 0.0
  
lmms_eval_specific_kwargs:
  default:
    prompt: "MULTI_CHOICE_DIRECT_PROMPT"