speecht5_improved_data_w_ljspeech

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: 0.5702

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 3000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.9234	0.8097	250	0.7694
0.7666	1.6194	500	0.6657
0.6923	2.4291	750	0.6291
0.6783	3.2389	1000	0.6109
0.6572	4.0486	1250	0.5961
0.6631	4.8583	1500	0.5909
0.6413	5.6680	1750	0.5809
0.6471	6.4777	2000	0.5798
0.6336	7.2874	2250	0.5733
0.6353	8.0972	2500	0.5723
0.6257	8.9069	2750	0.5700
0.632	9.7166	3000	0.5702

Framework versions

• Transformers 4.44.2
• Pytorch 2.1.2
• Datasets 2.17.0
• Tokenizers 0.19.1