donut-base-vishnu2

This model is a fine-tuned version of naver-clova-ix/donut-base on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.6822

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: 2e-05
• train_batch_size: 2
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
3.5134	1.0	137	2.8022
3.0637	2.0	274	1.8503
0.9705	3.0	411	1.6708
0.889	4.0	548	1.5365
0.3093	5.0	685	1.5766
0.1843	6.0	822	1.4999
0.7406	7.0	959	1.4946
2.8585	8.0	1096	1.5358
0.8517	9.0	1233	1.6042
0.7837	10.0	1370	1.6064
0.4483	11.0	1507	1.6338
0.5618	12.0	1644	1.6822

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1