speecht5_finetuned_voxpopuli_de_16

This model is a fine-tuned version of bregsi/speecht5_finetuned_voxpopuli_de on the voxpopuli dataset. It achieves the following results on the evaluation set:

• Loss: 0.4477

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: 4
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• 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: 12000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.4928	2.2812	1000	0.4584
0.4885	4.5623	2000	0.4555
0.4831	6.8435	3000	0.4523
0.4815	9.1246	4000	0.4515
0.4786	11.4058	5000	0.4508
0.4735	13.6869	6000	0.4491
0.4734	15.9681	7000	0.4494
0.4729	18.2492	8000	0.4482
0.4678	20.5304	9000	0.4483
0.4722	22.8115	10000	0.4479
0.47	25.0927	11000	0.4481
0.4682	27.3738	12000	0.4477

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.1+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1