eval_final_results.py

from compute_accuracy import compute_accuracy

def eval_final(test_metafile,dev_metafile,to_eval):
    print("Computing accuracy...")
    result_test = compute_accuracy(to_eval, test_metafile)
    result_dev = compute_accuracy(to_eval, dev_metafile)

    output = {"dev avg": result_dev['answered_acc'],
              "test avg": result_test['answered_acc'],
              **result_test}