vilt_finetuned_2 / README.md
farishehzad's picture
End of training
3e4917f verified
metadata
license: apache-2.0
base_model: dandelin/vilt-b32-mlm
tags:
  - generated_from_trainer
metrics:
  - f1
  - accuracy
model-index:
  - name: vilt_finetuned_2
    results: []

vilt_finetuned_2

This model is a fine-tuned version of dandelin/vilt-b32-mlm on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 3.3663
  • F1: 0.6000
  • Roc Auc: 0.7866
  • Accuracy: 0.5735

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: 5e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss F1 Roc Auc Accuracy
44.1455 1.0 129 6.5479 0.1270 0.5367 0.0735
2.9608 2.0 258 2.7634 0.4385 0.6965 0.3934
2.3046 3.0 387 2.4919 0.4948 0.7204 0.4412
1.895 4.0 516 2.3418 0.5652 0.7627 0.5257
1.4785 5.0 645 2.6462 0.5720 0.7701 0.5404
1.1491 6.0 774 2.8805 0.6074 0.7884 0.5772
0.8297 7.0 903 3.1832 0.5977 0.7866 0.5735
0.7249 8.0 1032 3.2679 0.6054 0.7903 0.5809
2.1554 9.0 1161 3.2926 0.6119 0.7940 0.5846
0.5323 10.0 1290 3.3663 0.6000 0.7866 0.5735

Framework versions

  • Transformers 4.40.0.dev0
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2