mms-1b-bem-female-sv

This model is a fine-tuned version of facebook/mms-1b-all on the BEMBASPEECH - BEM dataset. It achieves the following results on the evaluation set:

• Loss: 0.2132
• Wer: 0.3557

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: 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: 500
• num_epochs: 5.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.3992	200	0.3566	0.5019
No log	0.7984	400	0.2620	0.4029
1.7214	1.1976	600	0.2546	0.4100
1.7214	1.5968	800	0.2359	0.3965
0.2801	1.9960	1000	0.2322	0.3810
0.2801	2.3952	1200	0.2305	0.3746
0.2801	2.7944	1400	0.2258	0.3336
0.2528	3.1936	1600	0.2262	0.4309
0.2528	3.5928	1800	0.2164	0.3514
0.2351	3.9920	2000	0.2215	0.3894
0.2351	4.3912	2200	0.2165	0.3624
0.2351	4.7904	2400	0.2132	0.3557

Framework versions

• Transformers 4.43.0
• Pytorch 2.3.1+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1