speecht5_improved_data

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: 0.5781

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: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.9442	1.5326	250	0.7612
0.7905	3.0651	500	0.6773
0.7439	4.5977	750	0.6415
0.7208	6.1303	1000	0.6286
0.7108	7.6628	1250	0.6159
0.6941	9.1954	1500	0.6098
0.6869	10.7280	1750	0.5993
0.6572	12.2605	2000	0.5945
0.6671	13.7931	2250	0.5922
0.6566	15.3257	2500	0.5882
0.6551	16.8582	2750	0.5880
0.6564	18.3908	3000	0.5852
0.6458	19.9234	3250	0.5793
0.6457	21.4559	3500	0.5815
0.6473	22.9885	3750	0.5834
0.6532	24.5211	4000	0.5781

Framework versions

• Transformers 4.41.2
• Pytorch 2.1.2
• Datasets 2.17.0
• Tokenizers 0.19.1