Edit model card

Visualize in Weights & Biases

ViT_ASVspoof_DF

This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

  • Loss: 1.8822
  • Accuracy: 0.8934
  • F1: 0.8431
  • Precision: 0.7982
  • Recall: 0.8934
  • Test: 1
  • Auc Roc: 0.3976

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: 128
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 2
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy F1 Precision Recall Test Auc Roc
0.3293 0.1078 50 0.5369 0.8934 0.8431 0.7982 0.8934 1 0.4810
0.1251 0.2155 100 0.7074 0.8934 0.8431 0.7982 0.8934 1 0.5209
0.0671 0.3233 150 0.8683 0.8934 0.8431 0.7982 0.8934 1 0.5390
0.0463 0.4310 200 0.8867 0.8934 0.8431 0.7982 0.8934 1 0.5820
0.0365 0.5388 250 0.9675 0.8934 0.8431 0.7982 0.8934 1 0.6129
0.0332 0.6466 300 1.1225 0.8934 0.8431 0.7982 0.8934 1 0.5544
0.0788 0.7543 350 1.1081 0.8934 0.8431 0.7982 0.8934 1 0.5776
0.0425 0.8621 400 1.4392 0.8934 0.8431 0.7982 0.8934 1 0.5835
0.0566 0.9698 450 1.8030 0.8934 0.8431 0.7982 0.8934 1 0.5043
0.0821 1.0776 500 1.8901 0.8934 0.8431 0.7982 0.8934 1 0.6352
0.1122 1.1853 550 1.8085 0.8934 0.8431 0.7982 0.8934 1 0.3735
0.0446 1.2931 600 1.9759 0.8934 0.8431 0.7982 0.8934 1 0.3383
0.0342 1.4009 650 1.9482 0.8934 0.8431 0.7982 0.8934 1 0.4254
0.028 1.5086 700 1.9181 0.8934 0.8431 0.7982 0.8934 1 0.3508
0.0195 1.6164 750 1.9146 0.8934 0.8431 0.7982 0.8934 1 0.4860
0.0107 1.7241 800 1.8752 0.8934 0.8431 0.7982 0.8934 1 0.4285
0.0092 1.8319 850 1.8792 0.8934 0.8431 0.7982 0.8934 1 0.4012
0.0 1.9397 900 1.8822 0.8934 0.8431 0.7982 0.8934 1 0.3976

Framework versions

  • Transformers 4.42.3
  • Pytorch 2.1.2
  • Datasets 2.20.0
  • Tokenizers 0.19.1
Downloads last month
3
Safetensors
Model size
85.8M params
Tensor type
F32
·
Inference API
Unable to determine this model's library. Check the docs .

Model tree for Bisher/ViT_ASVspoof_DF

Finetuned
(502)
this model

Evaluation results