speecht5_kavinda4

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: 0.3649

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
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.7372	1.0	23	0.5522
0.592	2.0	46	0.5092
0.5346	3.0	69	0.4220
0.487	4.0	92	0.4280
0.4494	5.0	115	0.4160
0.4327	6.0	138	0.3919
0.4263	7.0	161	0.3925
0.4084	8.0	184	0.3778
0.4011	9.0	207	0.3755
0.4002	10.0	230	0.3725
0.3875	11.0	253	0.3659
0.3824	12.0	276	0.3705
0.381	13.0	299	0.3717
0.3792	14.0	322	0.3663
0.3842	15.0	345	0.3660
0.3714	16.0	368	0.3679
0.37	17.0	391	0.3621
0.3646	18.0	414	0.3693
0.3675	19.0	437	0.3620
0.3623	20.0	460	0.3649

Framework versions

• Transformers 4.53.1
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.2