wav2vec2-fleur-mms-batch8-epoch30-lr2e-4-ws200

This model is a fine-tuned version of Nachuwu/wav2vec2-fleur-mms-batch4-epoch12 on the audiofolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.0794
• Wer: 0.4891

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: 0.0002
• 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: 200
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
9.9931	3.85	100	7.4213	1.0117
3.6052	7.69	200	2.5721	0.9065
1.3783	11.54	300	1.3014	0.5148
1.0623	15.38	400	1.2017	0.5039
1.0293	19.23	500	1.1113	0.4938
1.0089	23.08	600	1.0877	0.4907
0.9897	26.92	700	1.0794	0.4891

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0