speecht5_finetuned_marar2000

This model is a fine-tuned version of maghrane/speecht5_finetuned_marar1000 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4801

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• 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
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.528	0.7976	100	0.5440
0.5453	1.5952	200	0.5655
0.5395	2.3928	300	0.5418
0.5421	3.1904	400	0.5570
0.5384	3.9880	500	0.5324
0.5447	4.7856	600	0.5454
0.5343	5.5833	700	0.5220
0.5165	6.3809	800	0.5269
0.5171	7.1785	900	0.5223
0.507	7.9761	1000	0.5227
0.5021	8.7737	1100	0.5052
0.4982	9.5713	1200	0.5081
0.488	10.3689	1300	0.5026
0.4787	11.1665	1400	0.4902
0.4849	11.9641	1500	0.4936
0.4778	12.7617	1600	0.4934
0.4646	13.5593	1700	0.4858
0.4615	14.3569	1800	0.4894
0.4716	15.1545	1900	0.4832
0.4607	15.9521	2000	0.4801

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Tokenizers 0.20.3