layoutlmv3-for-receipt-understanding

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

• Loss: 0.1719
• Precision: 0.9739
• Recall: 0.9837
• F1: 0.9788
• Accuracy: 0.9784

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: 5e-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	0.3125	50	1.0562	0.6706	0.7585	0.7118	0.7602
No log	0.625	100	0.6576	0.8260	0.8548	0.8401	0.8379
No log	0.9375	150	0.4929	0.8312	0.8758	0.8529	0.8693
No log	1.25	200	0.4159	0.8685	0.9076	0.8876	0.9028
No log	1.5625	250	0.3272	0.8998	0.9340	0.9166	0.9342
No log	1.875	300	0.3002	0.8953	0.9293	0.9120	0.9261
No log	2.1875	350	0.2275	0.9289	0.9542	0.9414	0.9423
No log	2.5	400	0.2756	0.9147	0.9402	0.9273	0.9317
No log	2.8125	450	0.2767	0.9312	0.9457	0.9384	0.9397
0.5451	3.125	500	0.2000	0.9416	0.9635	0.9524	0.9529
0.5451	3.4375	550	0.1736	0.9564	0.9705	0.9634	0.9648
0.5451	3.75	600	0.1865	0.9511	0.9674	0.9592	0.9643
0.5451	4.0625	650	0.1916	0.9511	0.9666	0.9588	0.9622
0.5451	4.375	700	0.1681	0.9519	0.9674	0.9596	0.9648
0.5451	4.6875	750	0.1808	0.9463	0.9705	0.9582	0.9660
0.5451	5.0	800	0.2651	0.9299	0.9581	0.9438	0.9406
0.5451	5.3125	850	0.2116	0.9497	0.9682	0.9589	0.9614
0.5451	5.625	900	0.2377	0.9481	0.9643	0.9561	0.9554
0.5451	5.9375	950	0.1865	0.9473	0.9627	0.9549	0.9614
0.1	6.25	1000	0.1875	0.9639	0.9736	0.9687	0.9686
0.1	6.5625	1050	0.1848	0.9542	0.9713	0.9627	0.9652
0.1	6.875	1100	0.2124	0.9548	0.9666	0.9606	0.9618
0.1	7.1875	1150	0.1733	0.9602	0.9744	0.9672	0.9690
0.1	7.5	1200	0.1955	0.9580	0.9728	0.9653	0.9631
0.1	7.8125	1250	0.1986	0.9505	0.9697	0.9600	0.9643
0.1	8.125	1300	0.1908	0.9579	0.9713	0.9645	0.9677
0.1	8.4375	1350	0.1689	0.9625	0.9752	0.9688	0.9694
0.1	8.75	1400	0.1836	0.9570	0.9682	0.9626	0.9665
0.1	9.0625	1450	0.1955	0.9639	0.9744	0.9691	0.9682
0.0396	9.375	1500	0.1577	0.9716	0.9814	0.9764	0.9758
0.0396	9.6875	1550	0.1689	0.9663	0.9790	0.9726	0.9745
0.0396	10.0	1600	0.1700	0.9693	0.9806	0.9749	0.9754
0.0396	10.3125	1650	0.1666	0.9708	0.9821	0.9765	0.9754
0.0396	10.625	1700	0.1713	0.9708	0.9798	0.9753	0.9745
0.0396	10.9375	1750	0.1892	0.9648	0.9775	0.9711	0.9716
0.0396	11.25	1800	0.1881	0.9693	0.9790	0.9741	0.9728
0.0396	11.5625	1850	0.1820	0.9625	0.9759	0.9692	0.9720
0.0396	11.875	1900	0.1757	0.9723	0.9821	0.9772	0.9750
0.0396	12.1875	1950	0.1772	0.9655	0.9767	0.9711	0.9720
0.0185	12.5	2000	0.1685	0.9701	0.9821	0.9761	0.9758
0.0185	12.8125	2050	0.1716	0.9716	0.9814	0.9764	0.9762
0.0185	13.125	2100	0.1697	0.9723	0.9814	0.9768	0.9762
0.0185	13.4375	2150	0.1704	0.9723	0.9821	0.9772	0.9771
0.0185	13.75	2200	0.1686	0.9739	0.9837	0.9788	0.9788
0.0185	14.0625	2250	0.1714	0.9724	0.9829	0.9776	0.9779
0.0185	14.375	2300	0.1709	0.9731	0.9837	0.9784	0.9788
0.0185	14.6875	2350	0.1713	0.9739	0.9837	0.9788	0.9784
0.0185	15.0	2400	0.1715	0.9739	0.9837	0.9788	0.9784
0.0185	15.3125	2450	0.1717	0.9739	0.9837	0.9788	0.9784
0.0083	15.625	2500	0.1719	0.9739	0.9837	0.9788	0.9784

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1