speecht5_tamil_2

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: 0.4098

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
0.4677	1.0	107	0.4364
0.4683	2.0	214	0.4379
0.4601	3.0	321	0.4262
0.4586	4.0	428	0.4258
0.4529	5.0	535	0.4237
0.4471	6.0	642	0.4234
0.446	7.0	749	0.4199
0.44	8.0	856	0.4195
0.4379	9.0	963	0.4185
0.4331	10.0	1070	0.4175
0.4307	11.0	1177	0.4169
0.427	12.0	1284	0.4140
0.4234	13.0	1391	0.4134
0.4225	14.0	1498	0.4117
0.4177	15.0	1605	0.4095
0.4158	16.0	1712	0.4098
0.4148	17.0	1819	0.4096
0.4123	18.0	1926	0.4095
0.4101	19.0	2033	0.4114
0.409	20.0	2140	0.4098

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.2