donut_experiment_bayesian_trial_12

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: 0.5083
• Bleu: 0.0675
• Precisions: [0.8421052631578947, 0.7822966507177034, 0.7423822714681441, 0.7006578947368421]
• Brevity Penalty: 0.0883
• Length Ratio: 0.2918
• Translation Length: 475
• Reference Length: 1628
• Cer: 0.7537
• Wer: 0.8211

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: 1.2643161326759464e-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: 3
• 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	0.4936	0.0660	[0.8375527426160337, 0.7673860911270983, 0.7277777777777777, 0.6897689768976898]	0.0876	0.2912	474	1628	0.7600	0.8274
0.0144	2.0	506	0.4987	0.0683	[0.8445378151260504, 0.7852028639618138, 0.7458563535911602, 0.7049180327868853]	0.0889	0.2924	476	1628	0.7515	0.8189
0.0089	3.0	759	0.5083	0.0675	[0.8421052631578947, 0.7822966507177034, 0.7423822714681441, 0.7006578947368421]	0.0883	0.2918	475	1628	0.7537	0.8211

Framework versions

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