bhojpuri_tts

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

• Loss: 0.4576

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: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 6000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.6508	2.5	200	0.5943
0.5656	5.0	400	0.5341
0.523	7.5	600	0.5141
0.512	10.0	800	0.4860
0.4966	12.5	1000	0.4803
0.522	15.0	1200	0.4921
0.4775	17.5	1400	0.4678
0.4726	20.0	1600	0.5031
0.4623	22.5	1800	0.4611
0.4612	25.0	2000	0.4593
0.4526	27.5	2200	0.4753
0.4558	30.0	2400	0.4578
0.4468	32.5	2600	0.4620
0.4474	35.0	2800	0.4618
0.4394	37.5	3000	0.4589
0.4332	40.0	3200	0.4463
0.4382	42.5	3400	0.4456
0.4382	45.0	3600	0.4481
0.4283	47.5	3800	0.4435
0.4278	50.0	4000	0.4470
0.4281	52.5	4200	0.4484
0.4236	55.0	4400	0.4482
0.422	57.5	4600	0.4480
0.4271	60.0	4800	0.4477
0.4105	62.5	5000	0.4475
0.4121	65.0	5200	0.4502
0.4115	67.5	5400	0.4522
0.4081	70.0	5600	0.4561
0.4059	72.5	5800	0.4610
0.4048	75.0	6000	0.4576

Framework versions

• Transformers 4.51.1
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1