layoutlmv3-large-cord

This model is a fine-tuned version of microsoft/layoutlmv3-large on the mp-02/cord dataset. It achieves the following results on the evaluation set:

• Loss: 0.1373
• Precision: 0.9705
• Recall: 0.9801
• F1: 0.9753
• Accuracy: 0.9739

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: 10
• eval_batch_size: 10
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.25	100	0.5589	0.8209	0.8583	0.8392	0.8394
No log	2.5	200	0.1936	0.9433	0.9644	0.9537	0.9579
No log	3.75	300	0.1456	0.9569	0.9760	0.9664	0.9698
No log	5.0	400	0.1368	0.9584	0.9743	0.9663	0.9726
0.4619	6.25	500	0.1448	0.9689	0.9809	0.9749	0.9744
0.4619	7.5	600	0.1286	0.9689	0.9818	0.9753	0.9753
0.4619	8.75	700	0.1311	0.9697	0.9809	0.9753	0.9748
0.4619	10.0	800	0.1335	0.9721	0.9809	0.9765	0.9758
0.4619	11.25	900	0.1355	0.9689	0.9793	0.9740	0.9753
0.0424	12.5	1000	0.1373	0.9705	0.9801	0.9753	0.9739

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu118
• Datasets 2.21.0
• Tokenizers 0.19.1