speecht5_finetuned_librispeech_polish_epo10_batch2_gas3

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: 0.3630

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 3
• total_train_batch_size: 6
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.4979	0.3836	100	0.4410
1.457	0.7673	200	0.4141
1.3563	1.1496	300	0.4106
1.2904	1.5332	400	0.3990
1.2615	1.9169	500	0.3980
1.2626	2.2992	600	0.3905
1.2478	2.6829	700	0.3859
1.2575	3.0652	800	0.3896
1.247	3.4488	900	0.3820
1.2218	3.8325	1000	0.3829
1.2306	4.2148	1100	0.3789
1.2151	4.5985	1200	0.3774
1.195	4.9821	1300	0.3760
1.1739	5.3645	1400	0.3752
1.1846	5.7481	1500	0.3721
1.1653	6.1304	1600	0.3739
1.1682	6.5141	1700	0.3689
1.1764	6.8977	1800	0.3699
1.1608	7.2801	1900	0.3690
1.1479	7.6637	2000	0.3680
1.1408	8.0460	2100	0.3667
1.133	8.4297	2200	0.3650
1.1269	8.8133	2300	0.3653
1.1187	9.1957	2400	0.3654
1.1193	9.5793	2500	0.3645
1.1246	9.9629	2600	0.3630

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0