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

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: 0.6594
• Bleu: 0.0554
• Precisions: [0.7702970297029703, 0.671875, 0.6061381074168798, 0.5538922155688623]
• Brevity Penalty: 0.0858
• Length Ratio: 0.2894
• Translation Length: 505
• Reference Length: 1745
• Cer: 0.7675
• Wer: 0.8512

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
3.4087	1.0	253	1.6485	0.0082	[0.4555984555984556, 0.14316702819956617, 0.06188118811881188, 0.01440922190201729]	0.0936	0.2968	518	1745	0.8353	0.9368
1.1999	2.0	506	0.8976	0.0274	[0.6748878923766816, 0.5347043701799485, 0.4578313253012048, 0.3890909090909091]	0.0543	0.2556	446	1745	0.7818	0.8754
0.8013	3.0	759	0.7295	0.0508	[0.7580645161290323, 0.6583143507972665, 0.5890052356020943, 0.5384615384615384]	0.0806	0.2842	496	1745	0.7629	0.8557
0.6404	4.0	1012	0.6594	0.0554	[0.7702970297029703, 0.671875, 0.6061381074168798, 0.5538922155688623]	0.0858	0.2894	505	1745	0.7675	0.8512

Framework versions

• Transformers 4.41.0.dev0
• Pytorch 2.1.0
• Datasets 2.19.1
• Tokenizers 0.19.1