speecht5_finetuned_nono

This model is a fine-tuned version of nonoJDWAOIDAWKDA/speecht5_finetuned_nono on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3932

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• 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: 1750
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.5131	29.6296	100	0.5124
0.4586	59.2593	200	0.5242
0.4348	88.8889	300	0.5369
0.4093	118.5185	400	0.5420
0.3839	148.1481	500	0.5394
0.3788	177.7778	600	0.5430
0.3686	207.4074	700	0.5504
0.3606	237.0370	800	0.5518
0.3555	266.6667	900	0.5524
0.3538	296.2963	1000	0.5475
0.4177	183.3333	1100	0.3818
0.3964	200.0	1200	0.3996
0.3819	216.6667	1300	0.3972
0.3782	233.3333	1400	0.3990
0.3746	250.0	1500	0.3984
0.3679	266.6667	1600	0.4062
0.3637	283.3333	1700	0.3932

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3