pseudolabeling-step1-F04

This model is a fine-tuned version of yongjian/wav2vec2-large-a on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5392
• Wer: 0.8870

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
21.4261	1.71	500	3.2064	1.0
2.9275	3.42	1000	2.6461	1.2637
2.49	5.14	1500	2.0627	1.2527
1.8582	6.85	2000	1.6367	1.1978
1.5071	8.56	2500	1.2845	1.1743
1.2181	10.27	3000	1.1395	1.1586
1.0386	11.99	3500	1.0155	1.0926
0.9307	13.7	4000	0.8144	1.0628
0.8073	15.41	4500	0.7666	1.1146
0.7209	17.12	5000	0.7020	1.0911
0.6618	18.84	5500	0.6829	1.0612
0.6079	20.55	6000	0.6023	0.9937
0.5242	22.26	6500	0.6057	0.9827
0.4848	23.97	7000	0.5802	0.9435
0.4602	25.68	7500	0.5376	0.9027
0.446	27.4	8000	0.5351	0.8964
0.4245	29.11	8500	0.5392	0.8870

Framework versions

• Transformers 4.23.1
• Pytorch 1.12.1+cu113
• Datasets 1.18.3
• Tokenizers 0.13.2