rop-bible-audio-aligned-speecht5

This model is a fine-tuned version of microsoft/speecht5_tts on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0511

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: 8
• eval_batch_size: 8
• seed: 3407
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 4000
• training_steps: 40000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.0787	1.1892	1000	0.0626
0.0745	2.3785	2000	0.0590
0.0701	3.5677	3000	0.0557
0.0689	4.7569	4000	0.0594
0.0667	5.9461	5000	0.0567
0.0664	7.1345	6000	0.0548
0.0644	8.3237	7000	0.0567
0.0631	9.5129	8000	0.0551
0.0614	10.7022	9000	0.0542
0.0607	11.8914	10000	0.0545
0.0612	13.0797	11000	0.0549
0.0582	14.2690	12000	0.0529
0.06	15.4582	13000	0.0542
0.059	16.6474	14000	0.0519
0.0563	17.8367	15000	0.0524
0.057	19.0250	16000	0.0519
0.0569	20.2142	17000	0.0523
0.0559	21.4035	18000	0.0513
0.0563	22.5927	19000	0.0517
0.0554	23.7819	20000	0.0512
0.055	24.9711	21000	0.0516
0.0537	26.1595	22000	0.0521
0.0537	27.3487	23000	0.0510
0.0544	28.5379	24000	0.0509
0.054	29.7272	25000	0.0512
0.052	30.9164	26000	0.0509
0.0517	32.1047	27000	0.0508
0.0522	33.2940	28000	0.0509
0.0516	34.4832	29000	0.0509
0.0518	35.6724	30000	0.0509
0.0503	36.8616	31000	0.0506
0.0516	38.0500	32000	0.0514
0.0514	39.2392	33000	0.0510
0.0524	40.4284	34000	0.0512
0.0497	41.6177	35000	0.0510
0.0508	42.8069	36000	0.0510
0.0512	43.9961	37000	0.0510
0.0491	45.1845	38000	0.0511
0.0512	46.3737	39000	0.0509
0.0504	47.5629	40000	0.0511

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.2