ceb_b32_le4_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: 0.3991

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: 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
0.4691	9.9010	500	0.4229
0.4352	19.8020	1000	0.4041
0.424	29.7030	1500	0.4032
0.4091	39.6040	2000	0.4037
0.4032	49.5050	2500	0.3995
0.3877	59.4059	3000	0.4005
0.3794	69.3069	3500	0.3979
0.3755	79.2079	4000	0.4074
0.3694	89.1089	4500	0.3979
0.3559	99.0099	5000	0.4004
0.3633	108.9109	5500	0.4046
0.3557	118.8119	6000	0.3987
0.3548	128.7129	6500	0.3986
0.3465	138.6139	7000	0.3982
0.3403	148.5149	7500	0.4006
0.34	158.4158	8000	0.3991

Framework versions

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