SpeechT5-Hausa-2

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: 0.5086

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: 200
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.5658	7.3733	200	0.5169
0.5266	14.7465	400	0.5300
0.4989	22.1198	600	0.4869
0.4747	29.4931	800	0.4763
0.4571	36.8664	1000	0.4736
0.4515	44.2396	1200	0.4751
0.4385	51.6129	1400	0.4884
0.4333	58.9862	1600	0.4969
0.429	66.3594	1800	0.5048
0.4198	73.7327	2000	0.5086

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1