diff --git "a/wandb/run-20250212_152709-lejyafmi/files/output.log" "b/wandb/run-20250212_152709-lejyafmi/files/output.log" --- "a/wandb/run-20250212_152709-lejyafmi/files/output.log" +++ "b/wandb/run-20250212_152709-lejyafmi/files/output.log" @@ -593,7 +593,1022 @@ warnings.warn( /home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.some-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. warnings.warn( - 39%|█████████████████████████████████████████████████▌ | 3096/8000 [1:09:39<1:34:36, 1.16s/it] + {'loss': 0.1758, 'grad_norm': 4.214677810668945, 'learning_rate': 6.637333333333333e-06, 'epoch': 0.38} {'loss': 0.1972, 'grad_norm': 4.144543647766113, 'learning_rate': 6.604000000000001e-06, 'epoch': 0.38} {'loss': 0.1293, 'grad_norm': 2.1775295734405518, 'learning_rate': 6.570666666666667e-06, 'epoch': 0.38} +{'loss': 0.099, 'grad_norm': 2.796152353286743, 'learning_rate': 6.537333333333334e-06, 'epoch': 0.39} +{'loss': 0.0945, 'grad_norm': 2.1920204162597656, 'learning_rate': 6.504e-06, 'epoch': 0.39} +{'loss': 0.1118, 'grad_norm': 2.8689582347869873, 'learning_rate': 6.470666666666667e-06, 'epoch': 0.39} +{'loss': 0.1732, 'grad_norm': 3.580993175506592, 'learning_rate': 6.4373333333333344e-06, 'epoch': 0.4} +{'loss': 0.1581, 'grad_norm': 3.9165573120117188, 'learning_rate': 6.404e-06, 'epoch': 0.4} +{'loss': 0.1716, 'grad_norm': 3.8235292434692383, 'learning_rate': 6.370666666666667e-06, 'epoch': 0.4} +{'loss': 0.1364, 'grad_norm': 3.21138072013855, 'learning_rate': 6.3373333333333345e-06, 'epoch': 0.41} +{'loss': 0.1459, 'grad_norm': 3.925539255142212, 'learning_rate': 6.304e-06, 'epoch': 0.41} +{'loss': 0.1668, 'grad_norm': 3.062764883041382, 'learning_rate': 6.270666666666667e-06, 'epoch': 0.41} +{'loss': 0.1243, 'grad_norm': 2.8379392623901367, 'learning_rate': 6.237333333333334e-06, 'epoch': 0.42} +{'loss': 0.0979, 'grad_norm': 2.979661226272583, 'learning_rate': 6.204e-06, 'epoch': 0.42} +{'loss': 0.0848, 'grad_norm': 2.4838883876800537, 'learning_rate': 6.170666666666667e-06, 'epoch': 0.42} +{'loss': 0.0927, 'grad_norm': 2.3293073177337646, 'learning_rate': 6.137333333333334e-06, 'epoch': 0.42} +{'loss': 0.0976, 'grad_norm': 3.3497400283813477, 'learning_rate': 6.104000000000001e-06, 'epoch': 0.43} +{'loss': 0.0881, 'grad_norm': 2.0302255153656006, 'learning_rate': 6.070666666666667e-06, 'epoch': 0.43} +{'loss': 0.0828, 'grad_norm': 2.112396001815796, 'learning_rate': 6.037333333333334e-06, 'epoch': 0.43} +{'loss': 0.0983, 'grad_norm': 2.513197183609009, 'learning_rate': 6.004000000000001e-06, 'epoch': 0.44} +{'loss': 0.0929, 'grad_norm': 2.1429622173309326, 'learning_rate': 5.970666666666667e-06, 'epoch': 0.44} +{'loss': 0.0916, 'grad_norm': 2.7300236225128174, 'learning_rate': 5.937333333333334e-06, 'epoch': 0.44} +{'loss': 0.1426, 'grad_norm': 4.011541366577148, 'learning_rate': 5.9040000000000006e-06, 'epoch': 0.45} +{'loss': 0.163, 'grad_norm': 3.1994545459747314, 'learning_rate': 5.870666666666667e-06, 'epoch': 0.45} +{'loss': 0.1568, 'grad_norm': 2.98388934135437, 'learning_rate': 5.837333333333333e-06, 'epoch': 0.45} +{'loss': 0.0937, 'grad_norm': 2.4515798091888428, 'learning_rate': 5.804000000000001e-06, 'epoch': 0.46} +{'loss': 0.0861, 'grad_norm': 2.0767834186553955, 'learning_rate': 5.770666666666666e-06, 'epoch': 0.46} +{'loss': 0.0917, 'grad_norm': 2.601104974746704, 'learning_rate': 5.7373333333333335e-06, 'epoch': 0.46} +{'loss': 0.1022, 'grad_norm': 2.593489408493042, 'learning_rate': 5.704000000000001e-06, 'epoch': 0.47} +{'loss': 0.1304, 'grad_norm': 3.5832834243774414, 'learning_rate': 5.670666666666668e-06, 'epoch': 0.47} +{'loss': 0.1634, 'grad_norm': 3.4403560161590576, 'learning_rate': 5.637333333333334e-06, 'epoch': 0.47} +{'loss': 0.1683, 'grad_norm': 3.6842737197875977, 'learning_rate': 5.604000000000001e-06, 'epoch': 0.47} +{'loss': 0.1538, 'grad_norm': 3.8382315635681152, 'learning_rate': 5.570666666666667e-06, 'epoch': 0.48} +{'loss': 0.165, 'grad_norm': 4.207257270812988, 'learning_rate': 5.537333333333334e-06, 'epoch': 0.48} +{'loss': 0.1558, 'grad_norm': 2.4130444526672363, 'learning_rate': 5.504e-06, 'epoch': 0.48} +{'loss': 0.1096, 'grad_norm': 2.3981151580810547, 'learning_rate': 5.4706666666666674e-06, 'epoch': 0.49} +{'loss': 0.0937, 'grad_norm': 2.2837915420532227, 'learning_rate': 5.437333333333333e-06, 'epoch': 0.49} +{'loss': 0.0876, 'grad_norm': 2.6647775173187256, 'learning_rate': 5.404e-06, 'epoch': 0.49} +{'loss': 0.15, 'grad_norm': 3.7677643299102783, 'learning_rate': 5.3706666666666675e-06, 'epoch': 0.5} +{'loss': 0.1925, 'grad_norm': 3.542175769805908, 'learning_rate': 5.337333333333333e-06, 'epoch': 0.5} +***** Running Evaluation ***** +[INFO|trainer.py:4180] 2025-02-12 16:54:50,654 >> Num examples: Unknown +[INFO|trainer.py:4181] 2025-02-12 16:54:50,654 >> Batch size = 16 +[INFO|trainer_utils.py:837] 2025-02-12 16:54:58,812 >> The following columns in the evaluation 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. +[INFO|generation_whisper.py:1844] 2025-02-12 16:54:58,999 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:00,211 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:01,715 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:03,007 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:04,217 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:05,579 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:06,663 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:07,716 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:08,870 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:09,965 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:11,072 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:12,108 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:13,097 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:14,014 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:15,076 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:15,989 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:16,893 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:17,929 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:18,887 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:19,797 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:20,802 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:21,787 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:22,749 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:23,731 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:24,709 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:25,688 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:26,745 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:27,889 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:28,825 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:29,817 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:30,877 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:31,914 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:32,973 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:34,005 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:35,132 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:36,186 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:37,147 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:38,209 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:39,239 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:40,203 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:41,211 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:42,259 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:43,227 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:44,221 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:45,189 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:46,268 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:47,252 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:48,255 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:49,402 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:50,418 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:51,459 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:52,515 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:53,447 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:54,476 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:55,511 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:56,620 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:57,540 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:58,475 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:55:59,523 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:00,465 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:01,524 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:02,521 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:03,583 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:04,595 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:05,503 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:06,411 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:07,417 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:08,417 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:09,431 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:10,421 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:11,400 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:12,471 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:13,455 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:14,528 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:15,525 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:16,543 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:17,471 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:18,467 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:19,471 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:20,442 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:21,387 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:22,327 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:23,330 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:24,372 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:25,405 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:26,355 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:27,363 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:28,457 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:29,472 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:30,595 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:31,624 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:32,612 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:33,582 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:34,550 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:35,646 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:36,674 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:37,687 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:38,697 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:39,660 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:40,685 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:41,723 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:42,793 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:43,792 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:44,867 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:45,989 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:47,016 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:48,077 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:49,183 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:50,171 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:51,120 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:52,155 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:53,192 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:54,144 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:55,101 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:56,092 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:57,046 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:58,118 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:56:59,145 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:00,139 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:01,144 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:02,173 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:03,111 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:04,149 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:05,210 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:06,183 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:07,142 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:08,122 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:09,142 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:10,120 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:11,120 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:12,097 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 16:57:13,020 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. + 50%|████████████████████████████████████████████████████████████████ | 4000/8000 [1:30:10<1:15:16, 1.13s/it][INFO|trainer.py:3860] 2025-02-12 16:57:21,319 >> Saving model checkpoint to ./checkpoint-4000 +{'eval_loss': 0.25648659467697144, 'eval_wer': 12.7372160418809, 'eval_runtime': 150.6646, 'eval_samples_per_second': 13.965, 'eval_steps_per_second': 0.876, 'epoch': 0.5} +[INFO|configuration_utils.py:423] 2025-02-12 16:57:21,320 >> Configuration saved in ./checkpoint-4000/config.json +[INFO|configuration_utils.py:906] 2025-02-12 16:57:21,321 >> Configuration saved in ./checkpoint-4000/generation_config.json +[INFO|modeling_utils.py:3040] 2025-02-12 16:57:22,731 >> Model weights saved in ./checkpoint-4000/model.safetensors +[INFO|feature_extraction_utils.py:437] 2025-02-12 16:57:22,732 >> Feature extractor saved in ./checkpoint-4000/preprocessor_config.json +[INFO|feature_extraction_utils.py:437] 2025-02-12 16:57:26,282 >> Feature extractor saved in ./preprocessor_config.json +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/datasets/formatting/torch_formatter.py:87: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). + return torch.tensor(value, **{**default_dtype, **self.torch_tensor_kwargs}) + 50%|███████████████████████████████████████████████████████████████▌ | 4001/8000 [1:30:19<54:04:14, 48.68s/it]It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. +02/12/2025 16:57:30 - WARNING - huggingface_hub.hf_api - It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. + 50%|████████████████████████████████████████████████████████████████▏ | 4008/8000 [1:30:28<5:43:05, 5.16s/it]/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.all-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.column-metadata-handling.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.some-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( + 57%|█████████████████████████████████████████████████████████████████████████▌ | 4600/8000 [1:43:49<1:03:53, 1.13s/it]'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: 415cf487-aa47-4982-8de5-bb40b3cb3a69)')' thrown while requesting GET https://huggingface.co./datasets/asierhv/composite_corpus_eu_v2.1/resolve/2c2153d993ce951528b5b62eb207632c6d267c29/data/train-00018-of-00150.parquet +{'loss': 0.1434, 'grad_norm': 2.5672571659088135, 'learning_rate': 5.304e-06, 'epoch': 0.5} +{'loss': 0.2075, 'grad_norm': 4.591808319091797, 'learning_rate': 5.270666666666668e-06, 'epoch': 0.51} +{'loss': 0.1478, 'grad_norm': 3.485185146331787, 'learning_rate': 5.237333333333334e-06, 'epoch': 0.51} +{'loss': 0.1383, 'grad_norm': 2.5995991230010986, 'learning_rate': 5.2040000000000005e-06, 'epoch': 0.51} +{'loss': 0.0959, 'grad_norm': 2.4682819843292236, 'learning_rate': 5.170666666666667e-06, 'epoch': 0.52} +{'loss': 0.0857, 'grad_norm': 2.436518669128418, 'learning_rate': 5.137333333333334e-06, 'epoch': 0.52} +{'loss': 0.0862, 'grad_norm': 2.0344107151031494, 'learning_rate': 5.104e-06, 'epoch': 0.52} +{'loss': 0.0808, 'grad_norm': 1.6771937608718872, 'learning_rate': 5.070666666666667e-06, 'epoch': 0.53} +{'loss': 0.0872, 'grad_norm': 1.7831439971923828, 'learning_rate': 5.037333333333334e-06, 'epoch': 0.53} +{'loss': 0.0832, 'grad_norm': 2.228795051574707, 'learning_rate': 5.004e-06, 'epoch': 0.53} +{'loss': 0.0927, 'grad_norm': 3.1402647495269775, 'learning_rate': 4.970666666666667e-06, 'epoch': 0.53} +{'loss': 0.1477, 'grad_norm': 3.662506580352783, 'learning_rate': 4.937333333333334e-06, 'epoch': 0.54} +{'loss': 0.1262, 'grad_norm': 2.865934371948242, 'learning_rate': 4.904000000000001e-06, 'epoch': 0.54} +{'loss': 0.1329, 'grad_norm': 3.2233200073242188, 'learning_rate': 4.870666666666667e-06, 'epoch': 0.54} +{'loss': 0.0795, 'grad_norm': 2.093703269958496, 'learning_rate': 4.837333333333334e-06, 'epoch': 0.55} +{'loss': 0.0715, 'grad_norm': 1.7601807117462158, 'learning_rate': 4.804e-06, 'epoch': 0.55} +{'loss': 0.0797, 'grad_norm': 2.1606643199920654, 'learning_rate': 4.770666666666667e-06, 'epoch': 0.55} +{'loss': 0.0883, 'grad_norm': 2.565343141555786, 'learning_rate': 4.737333333333334e-06, 'epoch': 0.56} +{'loss': 0.0965, 'grad_norm': 2.062619924545288, 'learning_rate': 4.704e-06, 'epoch': 0.56} +{'loss': 0.0891, 'grad_norm': 2.2219879627227783, 'learning_rate': 4.6706666666666675e-06, 'epoch': 0.56} +{'loss': 0.1147, 'grad_norm': 2.857029676437378, 'learning_rate': 4.637333333333334e-06, 'epoch': 0.57} +{'loss': 0.144, 'grad_norm': 3.090247392654419, 'learning_rate': 4.604e-06, 'epoch': 0.57} +{'loss': 0.1451, 'grad_norm': 3.8906264305114746, 'learning_rate': 4.570666666666667e-06, 'epoch': 0.57} +{'loss': 0.1475, 'grad_norm': 3.7733590602874756, 'learning_rate': 4.537333333333334e-06, 'epoch': 0.57} +02/12/2025 17:11:00 - WARNING - huggingface_hub.utils._http - '(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: 415cf487-aa47-4982-8de5-bb40b3cb3a69)')' thrown while requesting GET https://huggingface.co./datasets/asierhv/composite_corpus_eu_v2.1/resolve/2c2153d993ce951528b5b62eb207632c6d267c29/data/train-00018-of-00150.parquet +Retrying in 1s [Retry 1/5]. +02/12/2025 17:11:00 - WARNING - huggingface_hub.utils._http - Retrying in 1s [Retry 1/5]. + 62%|█████████████████████████████████████████████████████████████████████████████████▎ | 5000/8000 [1:52:57<58:08, 1.16s/it][INFO|trainer.py:4176] 2025-02-12 17:20:08,509 >> +{'loss': 0.1509, 'grad_norm': 3.379163980484009, 'learning_rate': 4.504e-06, 'epoch': 0.58} +{'loss': 0.1444, 'grad_norm': 3.4210824966430664, 'learning_rate': 4.470666666666667e-06, 'epoch': 0.58} +{'loss': 0.1295, 'grad_norm': 3.7809910774230957, 'learning_rate': 4.437333333333333e-06, 'epoch': 0.58} +{'loss': 0.1158, 'grad_norm': 2.537574052810669, 'learning_rate': 4.4040000000000005e-06, 'epoch': 0.59} +{'loss': 0.1249, 'grad_norm': 3.482285261154175, 'learning_rate': 4.370666666666667e-06, 'epoch': 0.59} +{'loss': 0.1238, 'grad_norm': 3.0114011764526367, 'learning_rate': 4.337333333333334e-06, 'epoch': 0.59} +{'loss': 0.0888, 'grad_norm': 2.117215394973755, 'learning_rate': 4.304000000000001e-06, 'epoch': 0.6} +{'loss': 0.0972, 'grad_norm': 2.0158379077911377, 'learning_rate': 4.270666666666667e-06, 'epoch': 0.6} +{'loss': 0.0793, 'grad_norm': 2.5208640098571777, 'learning_rate': 4.2373333333333335e-06, 'epoch': 0.6} +{'loss': 0.1035, 'grad_norm': 2.820002555847168, 'learning_rate': 4.204e-06, 'epoch': 0.61} +{'loss': 0.1128, 'grad_norm': 3.1144282817840576, 'learning_rate': 4.170666666666667e-06, 'epoch': 0.61} +{'loss': 0.1217, 'grad_norm': 3.1345527172088623, 'learning_rate': 4.137333333333334e-06, 'epoch': 0.61} +{'loss': 0.1061, 'grad_norm': 2.2702696323394775, 'learning_rate': 4.104e-06, 'epoch': 0.62} +{'loss': 0.0919, 'grad_norm': 2.714102268218994, 'learning_rate': 4.072e-06, 'epoch': 0.62} +{'loss': 0.0855, 'grad_norm': 2.448854923248291, 'learning_rate': 4.0386666666666666e-06, 'epoch': 0.62} +{'loss': 0.0818, 'grad_norm': 2.9392127990722656, 'learning_rate': 4.005333333333334e-06, 'epoch': 0.62} +***** Running Evaluation ***** +[INFO|trainer.py:4180] 2025-02-12 17:20:08,509 >> Num examples: Unknown +[INFO|trainer.py:4181] 2025-02-12 17:20:08,509 >> Batch size = 16 +[INFO|trainer_utils.py:837] 2025-02-12 17:20:24,466 >> The following columns in the evaluation 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. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:24,644 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:25,696 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:27,082 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:28,226 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:29,248 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:30,337 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:31,421 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:32,606 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:33,725 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:34,928 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:36,023 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:37,063 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:38,060 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:38,972 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:40,096 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:41,013 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:41,915 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:42,954 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:43,909 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:44,822 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:45,821 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:46,805 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:47,791 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:48,772 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:49,751 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:50,777 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:51,848 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:53,018 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:53,949 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:54,953 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:56,014 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:57,050 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:58,105 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:20:59,102 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:00,186 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:01,253 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:02,233 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:03,298 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:04,334 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:06,536 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:07,528 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:08,547 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:09,474 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:10,457 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:11,438 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:12,532 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:13,546 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:14,512 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:15,612 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:16,624 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:17,677 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:18,747 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:19,689 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:20,847 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:21,908 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:23,107 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:24,023 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:24,971 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:26,027 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:26,981 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:28,042 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:29,030 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:30,075 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:31,076 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:32,046 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:32,943 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:33,966 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:34,964 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:35,964 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:36,954 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:37,941 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:39,031 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:40,029 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:41,103 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:42,101 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:43,121 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:44,058 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:45,056 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:46,076 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:47,057 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:48,025 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:48,993 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:50,011 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:51,071 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:52,108 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:53,099 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:54,122 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:55,245 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:56,295 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:57,450 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:58,479 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:21:59,479 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:00,457 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:01,456 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:02,577 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:03,618 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:04,668 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:05,747 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:06,874 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:07,937 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:08,961 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:10,020 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:10,991 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:12,065 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:13,172 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:14,180 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:15,210 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:16,291 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:17,283 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:18,236 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:19,270 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:20,305 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:21,261 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:22,220 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:23,229 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:24,175 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:25,245 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:26,288 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:27,280 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:28,269 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:29,282 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:30,210 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:31,251 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:32,260 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:33,232 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:34,206 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:35,172 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:36,205 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:37,205 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:38,228 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:39,232 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:22:40,181 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. + 62%|█████████████████████████████████████████████████████████████████████████████████▎ | 5000/8000 [1:55:37<58:08, 1.16s/it][INFO|trainer.py:3860] 2025-02-12 17:22:48,522 >> Saving model checkpoint to ./checkpoint-5000 +{'eval_loss': 0.2562941014766693, 'eval_wer': 11.942600729176405, 'eval_runtime': 160.0125, 'eval_samples_per_second': 13.149, 'eval_steps_per_second': 0.825, 'epoch': 0.62} +[INFO|configuration_utils.py:423] 2025-02-12 17:22:48,523 >> Configuration saved in ./checkpoint-5000/config.json +[INFO|configuration_utils.py:906] 2025-02-12 17:22:48,524 >> Configuration saved in ./checkpoint-5000/generation_config.json +[INFO|modeling_utils.py:3040] 2025-02-12 17:22:49,924 >> Model weights saved in ./checkpoint-5000/model.safetensors +[INFO|feature_extraction_utils.py:437] 2025-02-12 17:22:49,926 >> Feature extractor saved in ./checkpoint-5000/preprocessor_config.json +[INFO|feature_extraction_utils.py:437] 2025-02-12 17:22:53,468 >> Feature extractor saved in ./preprocessor_config.json +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/datasets/formatting/torch_formatter.py:87: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). + return torch.tensor(value, **{**default_dtype, **self.torch_tensor_kwargs}) + 63%|███████████████████████████████████████████████████████████████████████████████▍ | 5001/8000 [1:55:46<42:48:38, 51.39s/it]It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. +02/12/2025 17:22:57 - WARNING - huggingface_hub.hf_api - It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. + 63%|████████████████████████████████████████████████████████████████████████████████ | 5006/8000 [1:55:53<8:05:45, 9.73s/it]/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.all-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.column-metadata-handling.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.some-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( + +{'loss': 0.1203, 'grad_norm': 2.4964210987091064, 'learning_rate': 3.972e-06, 'epoch': 0.63} +{'loss': 0.111, 'grad_norm': 3.330078125, 'learning_rate': 3.938666666666667e-06, 'epoch': 0.63} +{'loss': 0.164, 'grad_norm': 3.6872191429138184, 'learning_rate': 3.905333333333334e-06, 'epoch': 0.63} +{'loss': 0.1515, 'grad_norm': 3.728769063949585, 'learning_rate': 3.872e-06, 'epoch': 0.64} +{'loss': 0.1334, 'grad_norm': 3.4183156490325928, 'learning_rate': 3.838666666666667e-06, 'epoch': 0.64} +{'loss': 0.134, 'grad_norm': 3.4580440521240234, 'learning_rate': 3.8053333333333336e-06, 'epoch': 0.64} +{'loss': 0.1088, 'grad_norm': 2.2719855308532715, 'learning_rate': 3.772e-06, 'epoch': 0.65} +{'loss': 0.0724, 'grad_norm': 2.3186910152435303, 'learning_rate': 3.7386666666666673e-06, 'epoch': 0.65} +{'loss': 0.0759, 'grad_norm': 1.8175565004348755, 'learning_rate': 3.7053333333333337e-06, 'epoch': 0.65} +{'loss': 0.0813, 'grad_norm': 2.0874826908111572, 'learning_rate': 3.6720000000000006e-06, 'epoch': 0.66} +{'loss': 0.0824, 'grad_norm': 1.9950120449066162, 'learning_rate': 3.638666666666667e-06, 'epoch': 0.66} +{'loss': 0.0835, 'grad_norm': 2.6349194049835205, 'learning_rate': 3.6053333333333334e-06, 'epoch': 0.66} +{'loss': 0.0823, 'grad_norm': 2.7667415142059326, 'learning_rate': 3.5720000000000003e-06, 'epoch': 0.67} +{'loss': 0.1077, 'grad_norm': 3.617748260498047, 'learning_rate': 3.538666666666667e-06, 'epoch': 0.67} +{'loss': 0.1268, 'grad_norm': 3.2603073120117188, 'learning_rate': 3.5053333333333335e-06, 'epoch': 0.67} +{'loss': 0.1206, 'grad_norm': 2.9681355953216553, 'learning_rate': 3.4720000000000004e-06, 'epoch': 0.68} +{'loss': 0.1279, 'grad_norm': 4.156548500061035, 'learning_rate': 3.438666666666667e-06, 'epoch': 0.68} +{'loss': 0.1177, 'grad_norm': 3.2013888359069824, 'learning_rate': 3.4053333333333337e-06, 'epoch': 0.68} +{'loss': 0.0946, 'grad_norm': 3.299403190612793, 'learning_rate': 3.372e-06, 'epoch': 0.68} +{'loss': 0.0944, 'grad_norm': 2.39630389213562, 'learning_rate': 3.338666666666667e-06, 'epoch': 0.69} +{'loss': 0.1149, 'grad_norm': 3.7624928951263428, 'learning_rate': 3.3053333333333338e-06, 'epoch': 0.69} +{'loss': 0.1373, 'grad_norm': 3.3170886039733887, 'learning_rate': 3.272e-06, 'epoch': 0.69} +{'loss': 0.1056, 'grad_norm': 2.2296531200408936, 'learning_rate': 3.238666666666667e-06, 'epoch': 0.7} +{'loss': 0.0724, 'grad_norm': 1.8995999097824097, 'learning_rate': 3.2053333333333334e-06, 'epoch': 0.7} +{'loss': 0.0604, 'grad_norm': 2.3782520294189453, 'learning_rate': 3.172e-06, 'epoch': 0.7} +{'loss': 0.0581, 'grad_norm': 2.2558810710906982, 'learning_rate': 3.138666666666667e-06, 'epoch': 0.71} +{'loss': 0.0713, 'grad_norm': 2.4040448665618896, 'learning_rate': 3.1053333333333336e-06, 'epoch': 0.71} +{'loss': 0.0773, 'grad_norm': 2.5696732997894287, 'learning_rate': 3.072e-06, 'epoch': 0.71} +{'loss': 0.0765, 'grad_norm': 2.237166404724121, 'learning_rate': 3.038666666666667e-06, 'epoch': 0.72} +{'loss': 0.0779, 'grad_norm': 1.8783671855926514, 'learning_rate': 3.0053333333333332e-06, 'epoch': 0.72} +{'loss': 0.0751, 'grad_norm': 2.096334457397461, 'learning_rate': 2.9720000000000005e-06, 'epoch': 0.72} +{'loss': 0.0711, 'grad_norm': 2.0362164974212646, 'learning_rate': 2.938666666666667e-06, 'epoch': 0.72} +{'loss': 0.0635, 'grad_norm': 1.7136311531066895, 'learning_rate': 2.9053333333333334e-06, 'epoch': 0.73} +{'loss': 0.0698, 'grad_norm': 2.754848003387451, 'learning_rate': 2.872e-06, 'epoch': 0.73} +{'loss': 0.0741, 'grad_norm': 2.058065176010132, 'learning_rate': 2.8386666666666666e-06, 'epoch': 0.73} +{'loss': 0.0938, 'grad_norm': 3.0389583110809326, 'learning_rate': 2.805333333333334e-06, 'epoch': 0.74} +{'loss': 0.1387, 'grad_norm': 3.4811720848083496, 'learning_rate': 2.7720000000000003e-06, 'epoch': 0.74} +{'loss': 0.1283, 'grad_norm': 3.2388477325439453, 'learning_rate': 2.7386666666666667e-06, 'epoch': 0.74} +{'loss': 0.1073, 'grad_norm': 3.083925247192383, 'learning_rate': 2.7053333333333336e-06, 'epoch': 0.75} +{'loss': 0.1038, 'grad_norm': 2.6847918033599854, 'learning_rate': 2.672e-06, 'epoch': 0.75} +***** Running Evaluation ***** +[INFO|trainer.py:4180] 2025-02-12 17:43:07,939 >> Num examples: Unknown +[INFO|trainer.py:4181] 2025-02-12 17:43:07,939 >> Batch size = 16 +[INFO|trainer_utils.py:837] 2025-02-12 17:43:22,192 >> The following columns in the evaluation 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. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:22,365 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:23,580 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:24,998 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:26,225 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:27,276 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:28,426 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:29,638 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:30,792 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:31,992 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:33,202 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:34,361 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:35,413 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:36,412 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:37,319 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:38,399 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:39,321 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:40,237 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:41,294 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:42,267 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:43,182 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:44,176 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:45,180 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:46,166 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:47,176 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:48,155 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:49,156 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:50,230 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:51,385 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:52,331 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:53,343 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:54,425 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:55,475 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:56,547 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:57,541 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:58,648 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:43:59,691 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:00,661 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:01,761 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:02,783 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:03,752 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:04,742 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:05,767 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:06,710 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:07,708 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:08,679 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:09,757 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:10,772 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:11,747 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:12,871 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:13,884 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:14,922 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:15,966 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:16,906 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:17,924 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:18,944 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:20,063 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:21,004 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:21,957 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:23,014 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:23,958 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:25,029 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:26,028 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:27,079 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:28,098 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:29,042 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:29,943 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:30,984 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:32,029 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:33,081 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:34,111 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:35,117 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:36,225 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:37,255 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:38,335 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:39,362 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:40,399 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:41,342 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:42,334 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:43,368 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:44,346 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:45,347 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:46,329 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:47,362 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:48,419 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:49,467 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:50,453 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:51,480 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:52,610 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:53,644 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:54,788 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:55,817 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:56,829 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:57,839 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:58,844 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:44:59,969 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:01,027 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:02,055 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:03,076 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:04,056 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:05,098 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:06,126 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:07,194 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:08,191 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:09,283 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:10,424 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:11,448 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:12,488 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:13,598 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:14,587 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:15,548 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:16,593 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:17,638 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:18,603 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:19,640 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:20,634 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:21,593 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:22,663 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:23,696 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:24,697 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:25,694 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:26,730 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:27,669 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:28,722 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:29,737 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:30,742 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:31,713 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:32,686 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:33,710 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:34,712 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:35,718 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:36,698 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 17:45:37,644 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. + 75%|█████████████████████████████████████████████████████████████████████████████████████████████████▌ | 6000/8000 [2:18:35<38:52, 1.17s/it][INFO|trainer.py:3860] 2025-02-12 17:45:46,009 >> Saving model checkpoint to ./checkpoint-6000 +{'eval_loss': 0.23902159929275513, 'eval_wer': 11.07319809292325, 'eval_runtime': 158.0693, 'eval_samples_per_second': 13.311, 'eval_steps_per_second': 0.835, 'epoch': 0.75} +[INFO|configuration_utils.py:423] 2025-02-12 17:45:46,010 >> Configuration saved in ./checkpoint-6000/config.json +[INFO|configuration_utils.py:906] 2025-02-12 17:45:46,011 >> Configuration saved in ./checkpoint-6000/generation_config.json +[INFO|modeling_utils.py:3040] 2025-02-12 17:45:47,414 >> Model weights saved in ./checkpoint-6000/model.safetensors +[INFO|feature_extraction_utils.py:437] 2025-02-12 17:45:47,415 >> Feature extractor saved in ./checkpoint-6000/preprocessor_config.json +[INFO|feature_extraction_utils.py:437] 2025-02-12 17:45:50,918 >> Feature extractor saved in ./preprocessor_config.json +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/datasets/formatting/torch_formatter.py:87: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). + return torch.tensor(value, **{**default_dtype, **self.torch_tensor_kwargs}) + 75%|███████████████████████████████████████████████████████████████████████████████████████████████▎ | 6002/8000 [2:18:44<19:50:45, 35.76s/it]It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. +02/12/2025 17:45:56 - WARNING - huggingface_hub.hf_api - It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. + 75%|█████████████████████████████████████████████████████████████████████████████████████████████████▊ | 6017/8000 [2:19:02<44:13, 1.34s/it]/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.all-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.column-metadata-handling.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.some-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( + +{'loss': 0.0987, 'grad_norm': 2.7315189838409424, 'learning_rate': 2.6386666666666673e-06, 'epoch': 0.75} +{'loss': 0.0858, 'grad_norm': 2.3389735221862793, 'learning_rate': 2.6053333333333337e-06, 'epoch': 0.76} +{'loss': 0.0764, 'grad_norm': 1.982534646987915, 'learning_rate': 2.572e-06, 'epoch': 0.76} +{'loss': 0.0731, 'grad_norm': 1.9040074348449707, 'learning_rate': 2.538666666666667e-06, 'epoch': 0.76} +{'loss': 0.0758, 'grad_norm': 2.654710054397583, 'learning_rate': 2.5053333333333334e-06, 'epoch': 0.77} +{'loss': 0.0824, 'grad_norm': 2.6400296688079834, 'learning_rate': 2.4720000000000002e-06, 'epoch': 0.77} +{'loss': 0.0822, 'grad_norm': 7.269197463989258, 'learning_rate': 2.438666666666667e-06, 'epoch': 0.77} +{'loss': 0.0818, 'grad_norm': 2.363656520843506, 'learning_rate': 2.4053333333333335e-06, 'epoch': 0.78} +{'loss': 0.0768, 'grad_norm': 2.4660115242004395, 'learning_rate': 2.3720000000000003e-06, 'epoch': 0.78} +{'loss': 0.0783, 'grad_norm': 3.3116371631622314, 'learning_rate': 2.3386666666666668e-06, 'epoch': 0.78} +{'loss': 0.0899, 'grad_norm': 2.595853090286255, 'learning_rate': 2.3053333333333336e-06, 'epoch': 0.78} +{'loss': 0.0953, 'grad_norm': 2.709597587585449, 'learning_rate': 2.2720000000000004e-06, 'epoch': 0.79} +{'loss': 0.1249, 'grad_norm': 2.4446637630462646, 'learning_rate': 2.238666666666667e-06, 'epoch': 0.79} +{'loss': 0.1171, 'grad_norm': 3.4412341117858887, 'learning_rate': 2.2053333333333333e-06, 'epoch': 0.79} +{'loss': 0.1065, 'grad_norm': 2.2719008922576904, 'learning_rate': 2.172e-06, 'epoch': 0.8} +{'loss': 0.0872, 'grad_norm': 1.9873290061950684, 'learning_rate': 2.138666666666667e-06, 'epoch': 0.8} +{'loss': 0.0765, 'grad_norm': 2.487403392791748, 'learning_rate': 2.1053333333333334e-06, 'epoch': 0.8} +{'loss': 0.0736, 'grad_norm': 2.4424736499786377, 'learning_rate': 2.0720000000000002e-06, 'epoch': 0.81} +{'loss': 0.1064, 'grad_norm': 3.1507577896118164, 'learning_rate': 2.0386666666666667e-06, 'epoch': 0.81} +{'loss': 0.0993, 'grad_norm': 2.6285648345947266, 'learning_rate': 2.0053333333333335e-06, 'epoch': 0.81} +{'loss': 0.1299, 'grad_norm': 4.1934967041015625, 'learning_rate': 1.972e-06, 'epoch': 0.82} +{'loss': 0.1195, 'grad_norm': 3.031852960586548, 'learning_rate': 1.9386666666666668e-06, 'epoch': 0.82} +{'loss': 0.1197, 'grad_norm': 2.9288837909698486, 'learning_rate': 1.9053333333333334e-06, 'epoch': 0.82} +{'loss': 0.1127, 'grad_norm': 2.890054225921631, 'learning_rate': 1.8720000000000002e-06, 'epoch': 0.82} +{'loss': 0.1155, 'grad_norm': 3.130406618118286, 'learning_rate': 1.8386666666666669e-06, 'epoch': 0.83} +{'loss': 0.1291, 'grad_norm': 2.7169485092163086, 'learning_rate': 1.8053333333333333e-06, 'epoch': 0.83} +{'loss': 0.1097, 'grad_norm': 2.7390034198760986, 'learning_rate': 1.7720000000000001e-06, 'epoch': 0.83} +{'loss': 0.1022, 'grad_norm': 2.161604166030884, 'learning_rate': 1.7386666666666668e-06, 'epoch': 0.84} +{'loss': 0.0779, 'grad_norm': 2.210451126098633, 'learning_rate': 1.7053333333333336e-06, 'epoch': 0.84} +{'loss': 0.0728, 'grad_norm': 2.426438808441162, 'learning_rate': 1.672e-06, 'epoch': 0.84} +{'loss': 0.0859, 'grad_norm': 2.8744237422943115, 'learning_rate': 1.6386666666666667e-06, 'epoch': 0.85} +{'loss': 0.1496, 'grad_norm': 2.8165483474731445, 'learning_rate': 1.6053333333333335e-06, 'epoch': 0.85} +{'loss': 0.1343, 'grad_norm': 4.0077738761901855, 'learning_rate': 1.5720000000000002e-06, 'epoch': 0.85} +{'loss': 0.1397, 'grad_norm': 3.8011586666107178, 'learning_rate': 1.538666666666667e-06, 'epoch': 0.86} +{'loss': 0.1262, 'grad_norm': 2.7379047870635986, 'learning_rate': 1.5053333333333334e-06, 'epoch': 0.86} +{'loss': 0.1188, 'grad_norm': 3.250950574874878, 'learning_rate': 1.472e-06, 'epoch': 0.86} +{'loss': 0.1103, 'grad_norm': 2.782945156097412, 'learning_rate': 1.438666666666667e-06, 'epoch': 0.87} +{'loss': 0.1147, 'grad_norm': 3.08154034614563, 'learning_rate': 1.4053333333333335e-06, 'epoch': 0.87} +{'loss': 0.1332, 'grad_norm': 3.5768070220947266, 'learning_rate': 1.372e-06, 'epoch': 0.87} +{'loss': 0.1282, 'grad_norm': 3.155341863632202, 'learning_rate': 1.3386666666666668e-06, 'epoch': 0.88} +***** Running Evaluation ***** +[INFO|trainer.py:4180] 2025-02-12 18:05:21,315 >> Num examples: Unknown +[INFO|trainer.py:4181] 2025-02-12 18:05:21,315 >> Batch size = 16 +[INFO|trainer_utils.py:837] 2025-02-12 18:05:29,617 >> The following columns in the evaluation 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. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:29,786 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:31,140 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:32,626 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:34,102 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:35,233 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:36,498 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:37,606 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:38,714 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:39,829 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:40,947 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:41,951 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:43,000 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:43,995 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:46,219 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:47,289 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:48,188 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:49,089 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:50,124 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:51,096 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:52,004 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:53,010 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:54,011 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:54,997 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:56,008 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:56,994 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:57,994 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:05:59,069 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:00,233 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:01,182 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:02,235 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:03,336 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:04,394 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:05,477 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:06,488 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:07,602 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:08,666 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:09,635 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:10,772 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:11,802 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:14,040 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:15,034 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:16,090 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:17,033 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:18,017 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:18,994 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:20,075 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:21,080 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:22,056 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:23,182 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:24,195 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:25,267 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:26,311 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:27,249 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:28,267 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:29,277 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:30,376 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:31,303 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:32,249 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:33,311 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:34,267 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:35,325 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:36,313 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:37,365 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:38,382 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:39,310 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:40,206 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:41,232 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:42,247 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:43,299 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:44,301 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:45,300 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:46,406 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:47,435 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:48,507 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:49,528 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:50,565 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:51,520 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:52,499 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:53,512 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:54,484 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:55,473 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:56,427 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:57,442 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:58,505 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:06:59,556 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:00,536 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:01,550 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:02,709 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:03,769 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:04,936 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:05,980 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:07,015 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:08,004 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:08,995 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:10,103 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:11,162 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:12,199 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:13,229 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:14,216 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:15,248 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:16,269 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:17,337 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:18,322 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:19,425 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:20,559 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:21,607 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:22,656 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:23,769 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:24,754 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:25,710 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:26,754 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:27,805 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:28,761 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:29,703 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:30,724 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:31,673 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:32,752 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:33,806 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:34,818 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:35,824 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:36,851 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:37,809 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:38,853 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:39,863 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:40,840 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:41,815 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:42,803 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:43,835 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:44,832 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:45,856 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:46,820 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:07:47,748 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. + 88%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████▊ | 7000/8000 [2:40:45<18:54, 1.13s/it][INFO|trainer.py:3860] 2025-02-12 18:07:56,147 >> Saving model checkpoint to ./checkpoint-7000 +{'eval_loss': 0.23438745737075806, 'eval_wer': 11.391044218005048, 'eval_runtime': 154.8314, 'eval_samples_per_second': 13.589, 'eval_steps_per_second': 0.853, 'epoch': 0.88} +[INFO|configuration_utils.py:423] 2025-02-12 18:07:56,148 >> Configuration saved in ./checkpoint-7000/config.json +[INFO|configuration_utils.py:906] 2025-02-12 18:07:56,149 >> Configuration saved in ./checkpoint-7000/generation_config.json +[INFO|modeling_utils.py:3040] 2025-02-12 18:07:57,530 >> Model weights saved in ./checkpoint-7000/model.safetensors +[INFO|feature_extraction_utils.py:437] 2025-02-12 18:07:57,531 >> Feature extractor saved in ./checkpoint-7000/preprocessor_config.json +[INFO|feature_extraction_utils.py:437] 2025-02-12 18:08:00,757 >> Feature extractor saved in ./preprocessor_config.json +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/datasets/formatting/torch_formatter.py:87: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). + return torch.tensor(value, **{**default_dtype, **self.torch_tensor_kwargs}) + 88%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████▏ | 7001/8000 [2:40:54<13:49:55, 49.85s/it]It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. +02/12/2025 18:08:05 - WARNING - huggingface_hub.hf_api - It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. + 88%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████ | 7005/8000 [2:41:00<3:37:53, 13.14s/it]/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.all-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.column-metadata-handling.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.some-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( + +{'loss': 0.2406, 'grad_norm': 11.062019348144531, 'learning_rate': 1.308e-06, 'epoch': 0.88} +{'loss': 0.3469, 'grad_norm': 4.648179531097412, 'learning_rate': 1.2746666666666669e-06, 'epoch': 0.88} +{'loss': 0.3421, 'grad_norm': 4.388245105743408, 'learning_rate': 1.2413333333333335e-06, 'epoch': 0.88} +{'loss': 0.2847, 'grad_norm': 4.806427478790283, 'learning_rate': 1.2080000000000001e-06, 'epoch': 0.89} +{'loss': 0.1671, 'grad_norm': 3.0818049907684326, 'learning_rate': 1.1746666666666668e-06, 'epoch': 0.89} +{'loss': 0.1313, 'grad_norm': 4.117819309234619, 'learning_rate': 1.1413333333333334e-06, 'epoch': 0.89} +{'loss': 0.1177, 'grad_norm': 2.8558835983276367, 'learning_rate': 1.108e-06, 'epoch': 0.9} +{'loss': 0.0911, 'grad_norm': 3.0425021648406982, 'learning_rate': 1.0746666666666669e-06, 'epoch': 0.9} +{'loss': 0.0898, 'grad_norm': 2.6587588787078857, 'learning_rate': 1.0413333333333333e-06, 'epoch': 0.9} +{'loss': 0.0922, 'grad_norm': 1.7572664022445679, 'learning_rate': 1.0080000000000001e-06, 'epoch': 0.91} +{'loss': 0.0753, 'grad_norm': 2.00393009185791, 'learning_rate': 9.746666666666668e-07, 'epoch': 0.91} +{'loss': 0.0628, 'grad_norm': 1.845981478691101, 'learning_rate': 9.413333333333334e-07, 'epoch': 0.91} +{'loss': 0.0696, 'grad_norm': 2.008112907409668, 'learning_rate': 9.080000000000001e-07, 'epoch': 0.92} +{'loss': 0.0897, 'grad_norm': 2.837357759475708, 'learning_rate': 8.746666666666668e-07, 'epoch': 0.92} +{'loss': 0.1227, 'grad_norm': 2.4842417240142822, 'learning_rate': 8.413333333333334e-07, 'epoch': 0.92} +{'loss': 0.1012, 'grad_norm': 2.7866716384887695, 'learning_rate': 8.08e-07, 'epoch': 0.93} +{'loss': 0.1141, 'grad_norm': 2.1826930046081543, 'learning_rate': 7.746666666666668e-07, 'epoch': 0.93} +{'loss': 0.0754, 'grad_norm': 2.014090061187744, 'learning_rate': 7.413333333333333e-07, 'epoch': 0.93} +{'loss': 0.0736, 'grad_norm': 2.1539175510406494, 'learning_rate': 7.08e-07, 'epoch': 0.93} +{'loss': 0.0684, 'grad_norm': 2.712541341781616, 'learning_rate': 6.746666666666667e-07, 'epoch': 0.94} +{'loss': 0.1414, 'grad_norm': 3.281242847442627, 'learning_rate': 6.413333333333334e-07, 'epoch': 0.94} +{'loss': 0.1895, 'grad_norm': 4.088025093078613, 'learning_rate': 6.08e-07, 'epoch': 0.94} +{'loss': 0.222, 'grad_norm': 4.144560813903809, 'learning_rate': 5.746666666666667e-07, 'epoch': 0.95} +{'loss': 0.1349, 'grad_norm': 1.8468823432922363, 'learning_rate': 5.413333333333334e-07, 'epoch': 0.95} +{'loss': 0.0872, 'grad_norm': 2.5354621410369873, 'learning_rate': 5.08e-07, 'epoch': 0.95} +{'loss': 0.0725, 'grad_norm': 1.83882737159729, 'learning_rate': 4.746666666666667e-07, 'epoch': 0.96} +{'loss': 0.0988, 'grad_norm': 3.42556095123291, 'learning_rate': 4.413333333333333e-07, 'epoch': 0.96} +{'loss': 0.1166, 'grad_norm': 2.682558059692383, 'learning_rate': 4.0800000000000005e-07, 'epoch': 0.96} +{'loss': 0.1257, 'grad_norm': 3.2471797466278076, 'learning_rate': 3.7466666666666674e-07, 'epoch': 0.97} +{'loss': 0.1114, 'grad_norm': 2.4202020168304443, 'learning_rate': 3.4133333333333337e-07, 'epoch': 0.97} +{'loss': 0.0811, 'grad_norm': 2.8282711505889893, 'learning_rate': 3.0800000000000006e-07, 'epoch': 0.97} +{'loss': 0.104, 'grad_norm': 4.20676326751709, 'learning_rate': 2.746666666666667e-07, 'epoch': 0.97} +{'loss': 0.2773, 'grad_norm': 4.955998420715332, 'learning_rate': 2.413333333333333e-07, 'epoch': 0.98} +{'loss': 0.1105, 'grad_norm': 2.0168468952178955, 'learning_rate': 2.08e-07, 'epoch': 0.98} +{'loss': 0.0808, 'grad_norm': 1.6335862874984741, 'learning_rate': 1.7466666666666667e-07, 'epoch': 0.98} +{'loss': 0.0786, 'grad_norm': 2.269954204559326, 'learning_rate': 1.4133333333333333e-07, 'epoch': 0.99} +{'loss': 0.0801, 'grad_norm': 2.0813560485839844, 'learning_rate': 1.0800000000000001e-07, 'epoch': 0.99} +{'loss': 0.0687, 'grad_norm': 1.6093230247497559, 'learning_rate': 7.466666666666667e-08, 'epoch': 0.99} +{'loss': 0.0814, 'grad_norm': 1.730695366859436, 'learning_rate': 4.133333333333334e-08, 'epoch': 1.0} +02/12/2025 18:28:18 - WARNING - huggingface_hub.utils._http - '(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: 009f3961-f00b-4456-8f3b-f5beeed73ba9)')' thrown while requesting GET https://huggingface.co./datasets/asierhv/composite_corpus_eu_v2.1/resolve/2c2153d993ce951528b5b62eb207632c6d267c29/data/train-00011-of-00150.parquet +Retrying in 1s [Retry 1/5]. +02/12/2025 18:28:18 - WARNING - huggingface_hub.utils._http - Retrying in 1s [Retry 1/5]. +100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 8000/8000 [3:01:15<00:00, 1.12s/it][INFO|trainer.py:4176] 2025-02-12 18:28:26,590 >> +{'loss': 0.0959, 'grad_norm': 3.418311595916748, 'learning_rate': 8e-09, 'epoch': 1.0} +***** Running Evaluation ***** +[INFO|trainer.py:4180] 2025-02-12 18:28:26,591 >> Num examples: Unknown +[INFO|trainer.py:4181] 2025-02-12 18:28:26,591 >> Batch size = 16 +[INFO|trainer_utils.py:837] 2025-02-12 18:28:36,443 >> The following columns in the evaluation 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. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:36,636 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:37,929 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:39,629 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:40,907 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:42,147 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:43,290 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:44,383 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:45,515 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:46,714 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:47,807 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:48,974 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:50,030 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:51,028 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:51,940 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:53,011 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:53,928 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:54,824 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:55,861 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:56,821 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:57,733 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:58,729 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:28:59,711 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:00,686 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:01,667 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:02,636 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:03,638 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:04,705 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:05,860 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:06,791 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:07,792 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:08,866 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:09,912 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:10,983 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:11,974 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:13,064 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:14,105 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:15,079 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:16,146 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:17,157 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:19,369 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:20,358 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:21,395 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:22,338 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:23,321 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:24,296 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:25,367 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:26,378 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:27,341 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:28,464 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:29,477 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:30,504 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:31,539 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:32,463 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:33,471 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:34,477 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:35,563 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:36,481 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:37,440 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:38,514 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:39,450 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:40,506 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:41,508 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:42,545 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:43,575 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:44,506 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:45,387 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:46,407 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:47,411 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:48,425 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:49,418 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:50,409 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:51,497 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:52,490 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:53,574 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:54,590 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:55,627 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:56,552 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:57,550 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:58,569 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:29:59,542 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:00,532 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:01,497 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:02,513 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:03,573 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:04,628 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:05,616 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:06,649 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:07,783 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:08,824 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:09,984 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:11,030 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:12,070 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:13,037 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:14,025 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:15,144 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:16,199 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:17,239 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:18,260 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:19,243 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:20,274 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:21,301 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:22,390 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:23,378 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:24,445 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:25,581 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:26,601 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:27,726 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:28,833 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:29,822 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:30,792 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:31,830 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:32,877 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:33,824 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:34,763 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:35,764 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:36,727 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:37,789 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:38,822 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:39,825 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:40,838 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:41,863 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:42,813 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:43,852 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:44,875 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:45,853 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:46,820 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:47,809 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:48,833 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:49,830 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:50,844 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:51,814 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +[INFO|generation_whisper.py:1844] 2025-02-12 18:30:52,737 >> Increase max_length from 225 to 228 since input is conditioned on previous segment. +100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 8000/8000 [3:03:50<00:00, 1.12s/it][INFO|trainer.py:3860] 2025-02-12 18:31:01,025 >> Saving model checkpoint to ./checkpoint-8000 +{'eval_loss': 0.1835634410381317, 'eval_wer': 10.886229784051602, 'eval_runtime': 154.4338, 'eval_samples_per_second': 13.624, 'eval_steps_per_second': 0.855, 'epoch': 1.0} +[INFO|configuration_utils.py:423] 2025-02-12 18:31:01,026 >> Configuration saved in ./checkpoint-8000/config.json +[INFO|configuration_utils.py:906] 2025-02-12 18:31:01,027 >> Configuration saved in ./checkpoint-8000/generation_config.json +[INFO|modeling_utils.py:3040] 2025-02-12 18:31:02,413 >> Model weights saved in ./checkpoint-8000/model.safetensors +[INFO|feature_extraction_utils.py:437] 2025-02-12 18:31:02,414 >> Feature extractor saved in ./checkpoint-8000/preprocessor_config.json +[INFO|feature_extraction_utils.py:437] 2025-02-12 18:31:05,951 >> Feature extractor saved in ./preprocessor_config.json +[INFO|trainer.py:2621] 2025-02-12 18:31:05,974 >> + +Training completed. Do not forget to share your model on huggingface.co/models =) + + +[INFO|trainer.py:2859] 2025-02-12 18:31:05,974 >> Loading best model from ./checkpoint-8000 (score: 10.886229784051602). +[WARNING|trainer.py:2996] 2025-02-12 18:31:06,378 >> There were missing keys in the checkpoint model loaded: ['proj_out.weight']. +100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 8000/8000 [3:03:55<00:00, 1.38s/it] +{'train_runtime': 11036.9074, 'train_samples_per_second': 23.195, 'train_steps_per_second': 0.725, 'train_loss': 0.17036041705310345, 'epoch': 1.0} +[INFO|trainer.py:4641] 2025-02-12 18:31:06,419 >> Waiting for the current checkpoint push to be finished, this might take a couple of minutes. +It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. +02/12/2025 18:31:10 - WARNING - huggingface_hub.hf_api - It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.all-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.column-metadata-handling.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.some-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +[INFO|trainer.py:3860] 2025-02-12 18:32:59,287 >> Saving model checkpoint to ./ +[INFO|configuration_utils.py:423] 2025-02-12 18:32:59,288 >> Configuration saved in ./config.json +[INFO|configuration_utils.py:906] 2025-02-12 18:32:59,290 >> Configuration saved in ./generation_config.json +[INFO|modeling_utils.py:3040] 2025-02-12 18:33:01,543 >> Model weights saved in ./model.safetensors +[INFO|feature_extraction_utils.py:437] 2025-02-12 18:33:01,544 >> Feature extractor saved in ./preprocessor_config.json +[INFO|trainer.py:3860] 2025-02-12 18:33:01,545 >> Saving model checkpoint to ./ +[INFO|configuration_utils.py:423] 2025-02-12 18:33:01,546 >> Configuration saved in ./config.json +[INFO|configuration_utils.py:906] 2025-02-12 18:33:01,547 >> Configuration saved in ./generation_config.json +[INFO|modeling_utils.py:3040] 2025-02-12 18:33:03,594 >> Model weights saved in ./model.safetensors +[INFO|feature_extraction_utils.py:437] 2025-02-12 18:33:03,596 >> Feature extractor saved in ./preprocessor_config.json +[INFO|modelcard.py:449] 2025-02-12 18:33:03,826 >> Dropping the following result as it does not have all the necessary fields: +{'task': {'name': 'Automatic Speech Recognition', 'type': 'automatic-speech-recognition'}, 'metrics': [{'name': 'Wer', 'type': 'wer', 'value': 10.886229784051602}]} +It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. +02/12/2025 18:33:07 - WARNING - huggingface_hub.hf_api - It seems you are trying to upload a large folder at once. This might take some time and then fail if the folder is too large. For such cases, it is recommended to upload in smaller batches or to use `HfApi().upload_large_folder(...)`/`huggingface-cli upload-large-folder` instead. For more details, check out https://huggingface.co./docs/huggingface_hub/main/en/guides/upload#upload-a-large-folder. +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.all-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.column-metadata-handling.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +/home/tknika/xezpeleta/whisper/whisper-small-eu/.venv/lib/python3.12/site-packages/huggingface_hub/hf_api.py:3937: UserWarning: It seems that you are about to commit a data file (.venv/lib/python3.12/site-packages/pyarrow/tests/data/parquet/v0.7.1.some-named-index.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI. + warnings.warn( +run-lejyafmi.wandb: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4.62M/4.62M [00:01<00:00, 3.10MB/s]