layoutlmv3-finetuned-cord_100

This model is a fine-tuned version of microsoft/layoutlmv3-base on the cord-layoutlmv3 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1726
• Precision: 0.9561
• Recall: 0.9618
• F1: 0.9590
• Accuracy: 0.9703

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: 1e-05
• train_batch_size: 5
• eval_batch_size: 5
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 3000

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.56	250	1.0075	0.7597	0.8046	0.7815	0.8145
1.3907	3.12	500	0.5155	0.8388	0.8683	0.8533	0.8841
1.3907	4.69	750	0.3486	0.8917	0.9117	0.9016	0.9283
0.3755	6.25	1000	0.2722	0.9211	0.9356	0.9283	0.9435
0.3755	7.81	1250	0.2399	0.9356	0.9461	0.9408	0.9533
0.1857	9.38	1500	0.2170	0.9376	0.9454	0.9415	0.9542
0.1857	10.94	1750	0.1917	0.9510	0.9588	0.9549	0.9660
0.1236	12.5	2000	0.1821	0.9502	0.9573	0.9538	0.9652
0.1236	14.06	2250	0.1870	0.9538	0.9588	0.9563	0.9669
0.0858	15.62	2500	0.1741	0.9583	0.9633	0.9608	0.9711
0.0858	17.19	2750	0.1726	0.9561	0.9611	0.9586	0.9690
0.0708	18.75	3000	0.1726	0.9561	0.9618	0.9590	0.9703

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.1+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0