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: 0.2702
• Precision: 0.9350
• Recall: 0.9469
• F1: 0.9409
• Accuracy: 0.9474

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: 2500

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	4.17	250	1.0496	0.6714	0.7507	0.7088	0.7746
1.4245	8.33	500	0.5492	0.8401	0.8728	0.8561	0.8735
1.4245	12.5	750	0.3773	0.8934	0.9162	0.9047	0.9240
0.3461	16.67	1000	0.3212	0.9287	0.9364	0.9325	0.9380
0.3461	20.83	1250	0.2888	0.9276	0.9401	0.9338	0.9440
0.1502	25.0	1500	0.2749	0.9299	0.9431	0.9365	0.9474
0.1502	29.17	1750	0.2741	0.9321	0.9446	0.9383	0.9469
0.0866	33.33	2000	0.2715	0.9328	0.9454	0.9390	0.9465
0.0866	37.5	2250	0.2740	0.9314	0.9446	0.9379	0.9452
0.0635	41.67	2500	0.2702	0.9350	0.9469	0.9409	0.9474

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0+cu116
• Datasets 2.8.0
• Tokenizers 0.13.2