donut_experiment_bayesian_trial_3

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: 0.5840
• Bleu: 0.0667
• Precisions: [0.8136645962732919, 0.7347417840375586, 0.6829268292682927, 0.6346153846153846]
• Brevity Penalty: 0.0934
• Length Ratio: 0.2967
• Translation Length: 483
• Reference Length: 1628
• Cer: 0.7599
• Wer: 0.8328

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: 0.00017060423589132634
• 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
0.2012	1.0	253	0.6694	0.0547	[0.7680851063829788, 0.6682808716707022, 0.6067415730337079, 0.5484949832775919]	0.0851	0.2887	470	1628	0.7597	0.8411
0.127	2.0	506	0.6071	0.0638	[0.7818930041152263, 0.6876456876456877, 0.6370967741935484, 0.5841269841269842]	0.0954	0.2985	486	1628	0.7570	0.8360
0.0766	3.0	759	0.5786	0.0655	[0.8125, 0.735224586288416, 0.6885245901639344, 0.6407766990291263]	0.0915	0.2948	480	1628	0.7564	0.8319
0.0259	4.0	1012	0.5840	0.0667	[0.8136645962732919, 0.7347417840375586, 0.6829268292682927, 0.6346153846153846]	0.0934	0.2967	483	1628	0.7599	0.8328

Framework versions

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