layoutlmv3-large-model1-clinical-vs-specialized

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

• Loss: 0.1348
• Accuracy: 0.9771
• Macro Precision: 0.9771
• Macro Recall: 0.9771
• Macro F1: 0.9771
• Weighted F1: 0.9771
• Precision Specialized Diagnostics Procedures: 0.9776
• Recall Specialized Diagnostics Procedures: 0.9766
• F1 Specialized Diagnostics Procedures: 0.9771
• Precision Clinical Evaluation Progress: 0.9766
• Recall Clinical Evaluation Progress: 0.9776
• F1 Clinical Evaluation Progress: 0.9771

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: 16
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAFACTOR and the args are: No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Macro Precision	Macro Recall	Macro F1	Weighted F1	Precision Specialized Diagnostics Procedures	Recall Specialized Diagnostics Procedures	F1 Specialized Diagnostics Procedures	Precision Clinical Evaluation Progress	Recall Clinical Evaluation Progress	F1 Clinical Evaluation Progress
0.1797	0.2426	500	0.6136	0.8732	0.8942	0.8732	0.8715	0.8715	0.8033	0.9885	0.8863	0.9851	0.7579	0.8567
0.1666	0.4851	1000	0.2275	0.9558	0.9568	0.9558	0.9558	0.9558	0.9354	0.9793	0.9568	0.9783	0.9324	0.9548
0.1152	0.7277	1500	0.1561	0.9681	0.9683	0.9681	0.9681	0.9681	0.9772	0.9586	0.9678	0.9593	0.9776	0.9684
0.1771	0.9703	2000	0.1491	0.9643	0.9645	0.9643	0.9643	0.9643	0.9733	0.9547	0.9639	0.9556	0.9738	0.9646
0.1685	1.2125	2500	0.1408	0.9719	0.9719	0.9719	0.9719	0.9719	0.9722	0.9716	0.9719	0.9717	0.9722	0.9719
0.1446	1.4551	3000	0.1437	0.9725	0.9725	0.9725	0.9725	0.9725	0.9702	0.9749	0.9725	0.9748	0.9700	0.9724
0.1136	1.6976	3500	0.1414	0.9711	0.9712	0.9711	0.9711	0.9711	0.9768	0.9651	0.9709	0.9655	0.9771	0.9713
0.0852	1.9402	4000	0.1387	0.9733	0.9734	0.9733	0.9733	0.9733	0.9806	0.9656	0.9731	0.9662	0.9809	0.9735
0.0429	2.1824	4500	0.1323	0.9768	0.9768	0.9768	0.9768	0.9768	0.9787	0.9749	0.9768	0.9750	0.9787	0.9769
0.0717	2.4250	5000	0.1373	0.9766	0.9766	0.9766	0.9766	0.9766	0.975	0.9782	0.9766	0.9781	0.9749	0.9765
0.0784	2.6676	5500	0.1257	0.9785	0.9785	0.9785	0.9785	0.9785	0.9787	0.9782	0.9785	0.9782	0.9787	0.9785
0.1109	2.9101	6000	0.1348	0.9771	0.9771	0.9771	0.9771	0.9771	0.9776	0.9766	0.9771	0.9766	0.9776	0.9771

Framework versions

• Transformers 4.57.6
• Pytorch 2.10.0+cu128
• Tokenizers 0.22.2