SpeechT5-Hausa-7

This model is a fine-tuned version of microsoft/speecht5_tts on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4703

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: 4
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.6092	1.8476	100	0.5552
0.5638	3.6952	200	0.5097
0.5384	5.5427	300	0.5010
0.527	7.3903	400	0.4850
0.5157	9.2379	500	0.4830
0.5061	11.0855	600	0.4727
0.4955	12.9330	700	0.4773
0.4855	14.7806	800	0.4692
0.4801	16.6282	900	0.4651
0.4793	18.4758	1000	0.4623
0.4753	20.3233	1100	0.4708
0.4606	22.1709	1200	0.4668
0.4595	24.0185	1300	0.4622
0.4558	25.8661	1400	0.4628
0.4581	27.7136	1500	0.4628
0.4532	29.5612	1600	0.4665
0.4487	31.4088	1700	0.4652
0.4416	33.2564	1800	0.4661
0.4475	35.1039	1900	0.4677
0.4375	36.9515	2000	0.4703

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1