task: squadv2
dataset_path: squad_v2
output_type: greedy_until
training_split: train
validation_split: validation
doc_to_text: "Title: {{title}}\n\nBackground: {{context}}\n\nQuestion: {{question}}\n\n Answer:"
doc_to_target: "{% if answers.text| length > 0 %}{{answers.text}}{% else %}{{['unanswerable']}}{% endif %}"
target_delimiter: ""
should_decontaminate: true
doc_to_decontamination_query: context
process_results: !function utils.process_results
generation_kwargs:
  until:
    - "\n\n"
    - "\n"
  do_sample: false
  temperature: 0.0
filter_list:
  - name: remove_whitespace
    filter:
      - function: remove_whitespace
      - function: take_first
metric_list:
  - metric: exact
    aggregation: !function utils.exact
    higher_is_better: true
  # - metric: f1
  #   aggregation: mean
  #   higher_is_better: true
  # - metric: HasAns_exact
  #   aggregation: mean
  #   higher_is_better: true
  # - metric: HasAns_f1
  #   aggregation: mean
  #   higher_is_better: true
  # - metric: NoAns_exact
  #   aggregation: mean
  #   higher_is_better: true
  # - metric: NoAns_f1
  #   aggregation: mean
  #   higher_is_better: true
  # - metric: best_exact
  #   aggregation: mean
  #   higher_is_better: true
  # - metric: best_f1
  #   aggregation: mean
  #   higher_is_better: true