metadata

license: mit
base_model: microsoft/speecht5_tts
tags:
  - generated_from_trainer
datasets:
  - voxpopuli
model-index:
  - name: speecht5_finetuned_voxpopuli_nl
    results: []

speecht5_finetuned_voxpopuli_nl

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: 0.4591

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: 4
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: 500
training_steps: 4000

Training results

Training Loss	Epoch	Step	Validation Loss
0.7034	0.43	100	0.6427
0.6812	0.86	200	0.5998
0.5928	1.29	300	0.5292
0.5662	1.72	400	0.5095
0.5547	2.15	500	0.5024
0.5356	2.58	600	0.4929
0.532	3.01	700	0.4902
0.5312	3.44	800	0.4835
0.5171	3.87	900	0.4800
0.5178	4.3	1000	0.4780
0.5152	4.73	1100	0.4777
0.5064	5.16	1200	0.4737
0.5122	5.59	1300	0.4726
0.5042	6.02	1400	0.4700
0.5075	6.45	1500	0.4712
0.5057	6.88	1600	0.4688
0.5037	7.31	1700	0.4674
0.504	7.74	1800	0.4665
0.4977	8.17	1900	0.4652
0.4976	8.6	2000	0.4653
0.5014	9.03	2100	0.4650
0.4951	9.46	2200	0.4632
0.493	9.89	2300	0.4626
0.4983	10.32	2400	0.4628
0.4952	10.75	2500	0.4627
0.4961	11.18	2600	0.4616
0.4965	11.61	2700	0.4618
0.4895	12.04	2800	0.4615
0.4898	12.47	2900	0.4600
0.5008	12.9	3000	0.4614
0.4896	13.33	3100	0.4603
0.4957	13.76	3200	0.4612
0.4878	14.19	3300	0.4598
0.4923	14.62	3400	0.4594
0.4939	15.05	3500	0.4596
0.4828	15.48	3600	0.4591
0.4865	15.91	3700	0.4588
0.4999	16.34	3800	0.4600
0.4895	16.77	3900	0.4587
0.4929	17.2	4000	0.4591

Framework versions

Transformers 4.31.0.dev0
Pytorch 1.12.1+cu116
Datasets 2.4.0
Tokenizers 0.12.1