phase2

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: 0.4544

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 3407
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• 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: linear
• lr_scheduler_warmup_steps: 50
• training_steps: 20000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.5917	2.6823	1000	0.5957
0.5782	5.3627	2000	0.5542
0.5593	8.0430	3000	0.5318
0.4973	10.7253	4000	0.4915
0.4883	13.4056	5000	0.4819
0.4868	16.0860	6000	0.4706
0.5138	18.7683	7000	0.4689
0.4571	21.4486	8000	0.4650
0.4557	24.1289	9000	0.4675
0.5072	26.8113	10000	0.4631
0.492	29.4916	11000	0.4604
0.4535	32.1719	12000	0.4581
0.4668	34.8543	13000	0.4550
0.4825	37.5346	14000	0.4593
0.4551	40.2149	15000	0.4568
0.4285	42.8972	16000	0.4554
0.4383	45.5776	17000	0.4544
0.393	48.2579	18000	0.4529
0.4406	50.9402	19000	0.4570
0.4519	53.6206	20000	0.4544

Framework versions

• Transformers 4.55.2
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4