donut_experiment_bayesian_trial_14

This model is a fine-tuned version of naver-clova-ix/donut-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5314
• Bleu: 0.0658
• Precisions: [0.8375527426160337, 0.7721822541966427, 0.7222222222222222, 0.6798679867986799]
• Brevity Penalty: 0.0876
• Length Ratio: 0.2912
• Translation Length: 474
• Reference Length: 1628
• Cer: 0.7563
• Wer: 0.8271

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: 4.135841372163364e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 2
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Precisions	Brevity Penalty	Length Ratio	Translation Length	Reference Length	Cer	Wer
0.0251	1.0	253	0.5332	0.0676	[0.8308977035490606, 0.7677725118483413, 0.7205479452054795, 0.6688311688311688]	0.0908	0.2942	479	1628	0.7498	0.8243
0.0185	2.0	506	0.5314	0.0658	[0.8375527426160337, 0.7721822541966427, 0.7222222222222222, 0.6798679867986799]	0.0876	0.2912	474	1628	0.7563	0.8271

Framework versions

• Transformers 4.40.0
• Pytorch 2.1.0
• Datasets 2.18.0
• Tokenizers 0.19.1