fil_b64_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: 0.4246

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: 32
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• 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: 2000
• training_steps: 8000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.4811	22.2222	500	0.4381
0.4495	44.4444	1000	0.4216
0.4293	66.6667	1500	0.4446
0.4246	88.8889	2000	0.4177
0.4094	111.1111	2500	0.4179
0.3944	133.3333	3000	0.4232
0.3794	155.5556	3500	0.4190
0.3768	177.7778	4000	0.4187
0.3743	200.0	4500	0.4276
0.3598	222.2222	5000	0.4232
0.3634	244.4444	5500	0.4203
0.3558	266.6667	6000	0.4219
0.3502	288.8889	6500	0.4230
0.3529	311.1111	7000	0.4268
0.3447	333.3333	7500	0.4254
0.3371	355.5556	8000	0.4246

Framework versions

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