ceb_b32_le5_s8000

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: 0.3928

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: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• training_steps: 8000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.5667	9.9010	500	0.4776
0.4838	19.8020	1000	0.4321
0.4604	29.7030	1500	0.4157
0.4373	39.6040	2000	0.4034
0.4359	49.5050	2500	0.4006
0.4236	59.4059	3000	0.3975
0.4196	69.3069	3500	0.3956
0.4183	79.2079	4000	0.3938
0.4148	89.1089	4500	0.3941
0.4034	99.0099	5000	0.3930
0.4137	108.9109	5500	0.3955
0.4094	118.8119	6000	0.3924
0.4112	128.7129	6500	0.3917
0.4041	138.6139	7000	0.3923
0.3989	148.5149	7500	0.3927
0.3989	158.4158	8000	0.3928

Framework versions

• Transformers 4.41.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1