funasr/bin/asr_inference_paraformer_timestamp.py
@@ -410,7 +410,7 @@ results = speech2text(**batch) if len(results) < 1: hyp = Hypothesis(score=0.0, scores={}, states={}, yseq=[]) results = [[" ", ["<space>"], [2], hyp, 10, 6]] * nbest results = [[" ", ["sil"], [2], hyp, 10, 6]] * nbest time_end = time.time() forward_time = time_end - time_beg lfr_factor = results[0][-1]
