data/adversarial_qa_droberta_generate_question/stats.train.json

{
  "examples": 10000,
  "inputs_max_tokens": 586,
  "inputs_tokens": 2032929,
  "targets_max_tokens": 48,
  "targets_tokens": 130096
}