data/adversarial_qa_droberta_based_on/stats.train.json

{
  "examples": 10000,
  "inputs_max_tokens": 592,
  "inputs_tokens": 1975814,
  "targets_max_tokens": 212,
  "targets_tokens": 63677
}