donut_experiment_bayesian_trial_6

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: 0.5515
• Bleu: 0.0683
• Precisions: [0.8127572016460906, 0.7412587412587412, 0.6854838709677419, 0.638095238095238]
• Brevity Penalty: 0.0954
• Length Ratio: 0.2985
• Translation Length: 486
• Reference Length: 1628
• Cer: 0.7532
• Wer: 0.8274

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: 0.00016063260663724173
• 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: 4
• 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
0.3276	1.0	253	0.6672	0.0589	[0.76875, 0.6737588652482269, 0.6092896174863388, 0.5436893203883495]	0.0915	0.2948	480	1628	0.7586	0.8473
0.2008	2.0	506	0.5780	0.0662	[0.7905544147843943, 0.7069767441860465, 0.6595174262734584, 0.6107594936708861]	0.0960	0.2991	487	1628	0.7559	0.8374
0.1356	3.0	759	0.5355	0.0651	[0.8238993710691824, 0.7452380952380953, 0.6942148760330579, 0.6535947712418301]	0.0895	0.2930	477	1628	0.7580	0.8299
0.0394	4.0	1012	0.5515	0.0683	[0.8127572016460906, 0.7412587412587412, 0.6854838709677419, 0.638095238095238]	0.0954	0.2985	486	1628	0.7532	0.8274

Framework versions

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