| 0%| | 0/17200 [00:00<?, ?it/s]/home/patrick/python_bin/transformers/models/wav2vec2/modeling_wav2vec2.py:1107: UserWarning: __floordiv__ is deprecated, and its behavior will change in a future version of pytorch. It currently rounds toward 0 (like the 'trunc' function NOT 'floor'). This results in incorrect rounding for negative values. To keep the current behavior, use torch.div(a, b, rounding_mode='trunc'), or for actual floor division, use torch.div(a, b, rounding_mode='floor'). | |
| return (input_length - kernel_size) // stride + 1 | |
| 1%|▎ | 99/17200 [02:02<4:37:06, 1.03it/s] | |
| 1%|▌ | 199/17200 [04:06<4:58:27, 1.05s/it] | |
| 2%|▉ | 300/17200 [06:11<4:34:36, 1.03it/s] | |
| 2%|█▏ | 400/17200 [08:17<6:27:10, 1.38s/it] | |
| 3%|█▌ | 500/17200 [10:20<6:35:24, 1.42s/it]The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. | |
| ***** Running Evaluation ***** | |
| Num examples = 4620 | |
| Batch size = 8 | |
| {'loss': 3.3349, 'learning_rate': 1.8712499999999997e-05, 'epoch': 1.45} | |
| Configuration saved in ./checkpoint-500/config.json | |
| {'eval_loss': 3.2857718467712402, 'eval_wer': 1.0, 'eval_runtime': 141.2158, 'eval_samples_per_second': 32.716, 'eval_steps_per_second': 4.093, 'epoch': 1.45} | |
| Model weights saved in ./checkpoint-500/pytorch_model.bin | |
| Configuration saved in ./checkpoint-500/preprocessor_config.json | |