250119-centralized_learning

This model is a fine-tuned version of facebook/wav2vec2-base-960h on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3624.4978
• Wer: 1.0
• Cer: 0.9981

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3768.9383	1.0	2900	3624.4978	1.0	0.9981
3664.7982	2.0	5800	3623.8293	1.0	0.9981
3628.479	3.0	8700	3623.8401	1.0	0.9981
3629.406	4.0	11600	3623.7019	1.0	0.9981
3625.2683	5.0	14500	3623.6763	1.0	0.9981
3624.6528	6.0	17400	3623.6826	1.0	0.9981
3624.5862	7.0	20300	3623.6912	1.0	0.9981
3624.5	8.0	23200	3623.6826	1.0	0.9981
3623.9722	9.0	26100	3623.6606	1.0	0.9981
3625.5082	10.0	29000	3623.7019	1.0	0.9981
3624.1595	11.0	31900	3623.6472	1.0	0.9981
3625.175	12.0	34800	3623.6831	1.0	0.9981
3625.192	13.0	37700	3623.6731	1.0	0.9981
3623.9662	14.0	40600	3623.6777	1.0	0.9981
3623.9175	15.0	43500	3623.6653	1.0	0.9981
3623.9675	16.0	46400	3623.6497	1.0	0.9981
3624.471	17.0	49300	3623.6682	1.0	0.9981
3623.9983	18.0	52200	3623.6560	1.0	0.9981
3624.317	19.0	55100	3623.6587	1.0	0.9981
3623.9965	20.0	58000	3623.6743	1.0	0.9981
3623.9733	21.0	60900	3623.6799	1.0	0.9981
3623.9163	22.0	63800	3623.6455	1.0	0.9981
3623.817	23.0	66700	3623.6487	1.0	0.9981
3623.8545	24.0	69600	3623.6912	1.0	0.9981
3623.9572	25.0	72500	3623.6763	1.0	0.9981
3624.0855	26.0	75400	3623.6682	1.0	0.9981
3623.9163	27.0	78300	3623.6519	1.0	0.9981
3623.813	28.0	81200	3623.6863	1.0	0.9981
3624.0905	29.0	84100	3623.6797	1.0	0.9981
3623.847	30.0	87000	3623.6743	1.0	0.9981
3624.7505	31.0	89900	3623.6768	1.0	0.9981
3623.794	32.0	92800	3623.6716	1.0	0.9981
3623.775	33.0	95700	3623.6409	1.0	0.9981
3624.021	34.0	98600	3623.6431	1.0	0.9981
3623.7987	35.0	101500	3623.6394	1.0	0.9981
3623.8223	36.0	104400	3623.6831	1.0	0.9981
3623.979	37.0	107300	3623.6785	1.0	0.9981
3623.809	38.0	110200	3623.6826	1.0	0.9981
3623.857	39.0	113100	3623.6643	1.0	0.9981
3623.8252	40.0	116000	3623.6731	1.0	0.9981
3623.916	41.0	118900	3623.6838	1.0	0.9981
3623.9533	42.0	121800	3623.6694	1.0	0.9981
3618.18	43.0	124700	nan	1.0	0.9981
0.0	44.0	127600	nan	1.0	0.9981
0.0	45.0	130500	nan	1.0	0.9981
0.0	46.0	133400	nan	1.0	0.9981
0.0	47.0	136300	nan	1.0	0.9981
0.0	48.0	139200	nan	1.0	0.9981
0.0	49.0	142100	nan	1.0	0.9981
0.0	50.0	145000	nan	1.0	0.9981
0.0	51.0	147900	nan	1.0	0.9981
0.0	52.0	150800	nan	1.0	0.9981
0.0	53.0	153700	nan	1.0	0.9981
0.0	54.0	156600	nan	1.0	0.9981
0.0	55.0	159500	nan	1.0	0.9981
0.0	56.0	162400	nan	1.0	0.9981
0.0	57.0	165300	nan	1.0	0.9981
0.0	58.0	168200	nan	1.0	0.9981
0.0	59.0	171100	nan	1.0	0.9981
0.0	60.0	174000	nan	1.0	0.9981
0.0	61.0	176900	nan	1.0	0.9981
0.0	62.0	179800	nan	1.0	0.9981
0.0	63.0	182700	nan	1.0	0.9981
0.0	64.0	185600	nan	1.0	0.9981
0.0	65.0	188500	nan	1.0	0.9981
0.0	66.0	191400	nan	1.0	0.9981
0.0	67.0	194300	nan	1.0	0.9981
0.0	68.0	197200	nan	1.0	0.9981
0.0	69.0	200100	nan	1.0	0.9981
0.0	70.0	203000	nan	1.0	0.9981
0.0	71.0	205900	nan	1.0	0.9981
0.0	72.0	208800	nan	1.0	0.9981
0.0	73.0	211700	nan	1.0	0.9981
0.0	74.0	214600	nan	1.0	0.9981
0.0	75.0	217500	nan	1.0	0.9981
0.0	76.0	220400	nan	1.0	0.9981
0.0	77.0	223300	nan	1.0	0.9981
0.0	78.0	226200	nan	1.0	0.9981
0.0	79.0	229100	nan	1.0	0.9981
0.0	80.0	232000	nan	1.0	0.9981
0.0	81.0	234900	nan	1.0	0.9981
0.0	82.0	237800	nan	1.0	0.9981
0.0	83.0	240700	nan	1.0	0.9981
0.0	84.0	243600	nan	1.0	0.9981
0.0	85.0	246500	nan	1.0	0.9981
0.0	86.0	249400	nan	1.0	0.9981
0.0	87.0	252300	nan	1.0	0.9981
0.0	88.0	255200	nan	1.0	0.9981
0.0	89.0	258100	nan	1.0	0.9981
0.0	90.0	261000	nan	1.0	0.9981
0.0	91.0	263900	nan	1.0	0.9981
0.0	92.0	266800	nan	1.0	0.9981
0.0	93.0	269700	nan	1.0	0.9981
0.0	94.0	272600	nan	1.0	0.9981
0.0	95.0	275500	nan	1.0	0.9981
0.0	96.0	278400	nan	1.0	0.9981
0.0	97.0	281300	nan	1.0	0.9981
0.0	98.0	284200	nan	1.0	0.9981
0.0	99.0	287100	nan	1.0	0.9981
0.0	100.0	290000	nan	1.0	0.9981

Framework versions

• Transformers 4.43.3
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.19.1