RahelM/Amharic_tts_all

TTS_Amharic_all

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

• Loss: 0.3840

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 30000

Training results

Training Loss	Epoch	Step	Validation Loss
0.4741	2.53	1000	0.4312
0.4522	5.06	2000	0.4119
0.4417	7.59	3000	0.4040
0.4378	10.13	4000	0.3973
0.4287	12.66	5000	0.3943
0.4249	15.19	6000	0.3923
0.4153	17.72	7000	0.3900
0.4191	20.25	8000	0.3880
0.4151	22.78	9000	0.3877
0.414	25.32	10000	0.3863
0.4113	27.85	11000	0.3857
0.4122	30.38	12000	0.3849
0.411	32.91	13000	0.3835
0.4057	35.44	14000	0.3832
0.407	37.97	15000	0.3831
0.4041	40.51	16000	0.3839
0.4059	43.04	17000	0.3824
0.405	45.57	18000	0.3832
0.4048	48.1	19000	0.3833
0.4013	50.63	20000	0.3843
0.3999	53.16	21000	0.3833
0.4003	55.7	22000	0.3837
0.4019	58.23	23000	0.3828
0.397	60.76	24000	0.3834
0.4011	63.29	25000	0.3843
0.3997	65.82	26000	0.3832
0.3981	68.35	27000	0.3841
0.4012	70.89	28000	0.3842
0.4008	73.42	29000	0.3843
0.3971	75.95	30000	0.3840

Framework versions

• Transformers 4.38.1
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2