mascir_fr_wav2vec_version1000

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4677
• Wer: 0.37

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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
5.3896	4.0	500	3.0842	1.0
1.6969	8.0	1000	0.6327	0.5956
0.3056	12.0	1500	0.5024	0.49
0.1441	16.0	2000	0.5241	0.45
0.091	20.0	2500	0.4997	0.44
0.0676	24.0	3000	0.5173	0.4456
0.0603	28.0	3500	0.4487	0.4122
0.0378	32.0	4000	0.4554	0.3933
0.0328	36.0	4500	0.4395	0.3822
0.0275	40.0	5000	0.4910	0.3889
0.0198	44.0	5500	0.4861	0.3722
0.019	48.0	6000	0.4677	0.37

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.3
• Tokenizers 0.13.3