speecht5_finetuned_voxpopuli_it

This model is a fine-tuned version of microsoft/speecht5_tts on the voxpopuli dataset. It achieves the following results on the evaluation set:

• Loss: 0.4968

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: 0.0005
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss
0.6707	1.0	108	0.5946
0.6625	2.0	217	0.6029
0.708	3.0	325	0.6118
0.6588	4.0	434	0.7109
0.6614	5.0	542	0.5799
0.6375	6.0	651	0.5714
0.619	7.0	759	0.5699
0.5806	8.0	868	0.5538
0.6024	9.0	976	0.5856
0.5728	10.0	1085	0.5446
0.5624	11.0	1193	0.5508
0.5711	12.0	1302	0.5376
0.5438	13.0	1410	0.5300
0.5308	14.0	1519	0.5206
0.5536	15.0	1627	0.5359
0.5285	16.0	1736	0.5264
0.525	17.0	1844	0.5108
0.4961	18.0	1953	0.5116
0.5111	19.0	2061	0.5042
0.4869	20.0	2170	0.5050
0.4864	21.0	2278	0.4994
0.4794	22.0	2387	0.5039
0.4787	23.0	2495	0.4975
0.4692	24.0	2604	0.4961
0.4656	24.88	2700	0.4968

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.0
• Datasets 2.1.0
• Tokenizers 0.13.3