aam-len3-bs256-lr1e-3

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

• Loss: 0.6434
• Accuracy: 0.9617

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: 0.001
• train_batch_size: 256
• eval_batch_size: 1
• seed: 914
• 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_ratio: 0.1
• num_epochs: 10.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
7.6771	1.0	523	6.7010	0.6003
3.5879	2.0	1046	1.9993	0.9141
1.9536	3.0	1569	0.8234	0.9607
1.5008	4.0	2092	0.6434	0.9617
0.0	5.0	2615	nan	0.0005
0.0	6.0	3138	nan	0.0005
0.0	7.0	3661	nan	0.0005
0.0	8.0	4184	nan	0.0005
0.0	9.0	4707	nan	0.0005
0.0	10.0	5230	nan	0.0005

Framework versions

• Transformers 4.48.3
• Pytorch 2.0.0+cu117
• Datasets 3.2.0
• Tokenizers 0.21.0