README.md

metadata

license: afl-3.0
datasets:
  - copenlu/fever_gold_evidence
language:
  - en
metrics:
  - precision
  - recall
  - f1

wandb: Run summary:
wandb:                 eval/f1 0.8823
wandb:               eval/loss 0.55886
wandb:                  eval/p 0.88088
wandb:                  eval/r 0.88558

Note:

1. [evidence_text][SEP][claim]
2. Only trained/validated on instances length <= 512 tokens.