ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53

This model is a fine-tuned version of gary109/ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53 on the GARY109/AI_LIGHT_DANCE - ONSET-SINGING3 dataset. It achieves the following results on the evaluation set:

• Loss: 0.8797
• Wer: 0.5513

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: 5e-06
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 30.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.9613	1.0	2309	1.0171	0.7271
0.8254	2.0	4618	0.9771	0.6650
0.7406	3.0	6927	0.9174	0.6420
0.74	4.0	9236	0.9551	0.6371
0.5855	5.0	11545	0.9262	0.6453
0.5536	6.0	13854	0.9056	0.5894
0.505	7.0	16163	0.9166	0.6029
0.449	8.0	18472	0.8816	0.5873
0.4219	9.0	20781	0.8970	0.5589
0.5764	10.0	23090	0.9189	0.5649
0.5075	11.0	25399	0.8797	0.5513
0.4366	12.0	27708	0.9011	0.5567
0.4915	13.0	30017	0.9248	0.5455
0.3554	14.0	32326	0.9309	0.5374
0.3975	15.0	34635	0.9103	0.5259
0.4119	16.0	36944	0.9402	0.5290
0.267	17.0	39253	0.9479	0.5115
0.3107	18.0	41562	0.9428	0.5099
0.2684	19.0	43871	0.9508	0.5133
0.2125	20.0	46180	0.9737	0.5097
0.3149	21.0	48489	0.9992	0.5095
0.2313	22.0	50798	1.0037	0.5059
0.2674	23.0	53107	1.0091	0.5040
0.2056	24.0	55416	1.0082	0.5076
0.2781	25.0	57725	1.0160	0.5015
0.2005	26.0	60034	1.0390	0.5131
0.2221	27.0	62343	1.0401	0.5074
0.1857	28.0	64652	1.0484	0.5096
0.1562	29.0	66961	1.0516	0.5064
0.3027	30.0	69270	1.0543	0.5049

Framework versions

• Transformers 4.21.0.dev0
• Pytorch 1.9.1+cu102
• Datasets 2.3.3.dev0
• Tokenizers 0.12.1