Reading metadata...: 2165it [00:00, 12846.67it/s] | 0/30000 [00:00> The following columns in the training set don't have a corresponding argument in `WhisperForConditionalGeneration.forward` and have been ignored: input_length. If input_length are not expected by `WhisperForConditionalGeneration.forward`, you can safely ignore this message. [WARNING|logging.py:329] 2023-11-18 12:34:22,289 >> `use_cache = True` is incompatible with gradient checkpointing. Setting `use_cache = False`... 0%| | 12/30000 [04:17<49:00:59, 5.88s/it] 0%| | 13/30000 [04:19<40:29:29, 4.86s/it] 0%|▏ | 25/30000 [05:27<45:57:37, 5.52s/it]