speecht5_finetuned_k_voice_v3

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: 0.4215

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-06
• train_batch_size: 2
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 200
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
0.6081	0.2579	500	0.5853
0.5601	0.5159	1000	0.4924
0.5331	0.7738	1500	0.4709
0.5263	1.0317	2000	0.4698
0.5151	1.2897	2500	0.4615
0.5172	1.5476	3000	0.4618
0.5109	1.8055	3500	0.4486
0.5025	2.0635	4000	0.4461
0.4822	2.3214	4500	0.4356
0.4911	2.5793	5000	0.4439
0.4931	2.8372	5500	0.4331
0.4904	3.0952	6000	0.4304
0.474	3.3531	6500	0.4330
0.4716	3.6110	7000	0.4307
0.4673	3.8690	7500	0.4274
0.4654	4.1269	8000	0.4250
0.4609	4.3848	8500	0.4215

Framework versions

• Transformers 4.40.1
• Pytorch 2.6.0
• Datasets 3.5.0
• Tokenizers 0.19.1