The switch-base-32 model was fine-tuned on the SQuAD dataset.
Validation exact-match/F1-score: 65.39/85.81.
The prompt key in PromptSource: "answer_given_context_and_question".
The switch-base-32 model was fine-tuned on the SQuAD dataset.
Validation exact-match/F1-score: 65.39/85.81.
The prompt key in PromptSource: "answer_given_context_and_question".