donut-base-stnk-no-address-v3

This model is a fine-tuned version of squantumengine/donut-base-stnk-no-address-v2 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2116

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: 5

Training results

Training Loss	Epoch	Step	Validation Loss
0.4954	1.0	3253	0.3898
0.3524	2.0	6506	0.2783
0.2498	3.0	9759	0.2342
0.1684	4.0	13012	0.2177
0.1442	5.0	16265	0.2116

Framework versions

• Transformers 4.31.0
• Pytorch 1.13.1+cu117
• Datasets 2.14.0
• Tokenizers 0.13.3