donut-base-sroie-metrics-combined-new-instance-050824-NUM-01

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: 0.4660
• Bleu: 0.0647
• Precisions: [0.8446215139442231, 0.7842696629213484, 0.7422680412371134, 0.7129909365558912]
• Brevity Penalty: 0.0841
• Length Ratio: 0.2877
• Translation Length: 502
• Reference Length: 1745
• Cer: 0.7571
• Wer: 0.8220

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: 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
0.8919	1.0	253	0.6234	0.0484	[0.7540983606557377, 0.6450116009280742, 0.5962566844919787, 0.5646687697160884]	0.0761	0.2797	488	1745	0.7741	0.8528
0.4361	2.0	506	0.5208	0.0584	[0.8161616161616162, 0.7442922374429224, 0.7007874015748031, 0.6635802469135802]	0.0800	0.2837	495	1745	0.7647	0.8335
0.2804	3.0	759	0.4764	0.0608	[0.8481781376518218, 0.7803203661327232, 0.7394736842105263, 0.6996904024767802]	0.0795	0.2831	494	1745	0.7568	0.8211
0.2261	4.0	1012	0.4660	0.0647	[0.8446215139442231, 0.7842696629213484, 0.7422680412371134, 0.7129909365558912]	0.0841	0.2877	502	1745	0.7571	0.8220

Framework versions

• Transformers 4.40.2
• Pytorch 2.1.0
• Datasets 2.19.1
• Tokenizers 0.19.1