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: 0.1929
• Precision: 0.9625
• Recall: 0.9759
• F1: 0.9692
• Accuracy: 0.9711

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
2.1555	0.3125	50	1.2193	0.5472	0.6801	0.6064	0.7012
1.0163	0.625	100	0.7732	0.7762	0.8323	0.8033	0.8145
0.7451	0.9375	150	0.5322	0.8109	0.8657	0.8374	0.8553
0.5237	1.25	200	0.4384	0.8228	0.8905	0.8553	0.8960
0.4004	1.5625	250	0.3511	0.8772	0.9154	0.8959	0.9160
0.3131	1.875	300	0.3527	0.8817	0.9255	0.9030	0.9138
0.2565	2.1875	350	0.2919	0.9142	0.9511	0.9323	0.9334
0.2381	2.5	400	0.2392	0.9218	0.9519	0.9366	0.9419
0.2053	2.8125	450	0.2374	0.9189	0.9495	0.9339	0.9419
0.1541	3.125	500	0.2738	0.9250	0.9581	0.9413	0.9385
0.1557	3.4375	550	0.2274	0.9253	0.9519	0.9384	0.9491
0.1534	3.75	600	0.2394	0.9445	0.9643	0.9543	0.9533
0.1045	4.0625	650	0.2286	0.9425	0.9666	0.9544	0.9567
0.1368	4.375	700	0.2270	0.9452	0.9643	0.9547	0.9516
0.0747	4.6875	750	0.2470	0.9377	0.9589	0.9482	0.9495
0.1099	5.0	800	0.2024	0.9467	0.9651	0.9558	0.9580
0.0575	5.3125	850	0.2077	0.9534	0.9697	0.9615	0.9652
0.0848	5.625	900	0.2099	0.9464	0.9596	0.9530	0.9559
0.0913	5.9375	950	0.2399	0.9428	0.9604	0.9515	0.9482
0.0482	6.25	1000	0.2054	0.9548	0.9682	0.9614	0.9622
0.0793	6.5625	1050	0.2136	0.9579	0.9720	0.9649	0.9567
0.0425	6.875	1100	0.2274	0.9535	0.9713	0.9623	0.9635
0.0442	7.1875	1150	0.2033	0.9509	0.9627	0.9568	0.9597
0.0425	7.5	1200	0.1676	0.9588	0.9759	0.9673	0.9682
0.0425	7.8125	1250	0.1998	0.9536	0.9728	0.9631	0.9648
0.0455	8.125	1300	0.1987	0.9639	0.9752	0.9695	0.9686
0.0246	8.4375	1350	0.2094	0.9594	0.9713	0.9653	0.9626
0.0502	8.75	1400	0.2046	0.9594	0.9720	0.9657	0.9605
0.028	9.0625	1450	0.1927	0.9573	0.9752	0.9662	0.9648
0.0223	9.375	1500	0.1728	0.9602	0.9752	0.9676	0.9703
0.0346	9.6875	1550	0.2364	0.9571	0.9705	0.9638	0.9635
0.0152	10.0	1600	0.1953	0.9610	0.9759	0.9684	0.9703
0.0176	10.3125	1650	0.2045	0.9632	0.9744	0.9687	0.9682
0.019	10.625	1700	0.2316	0.9587	0.9728	0.9657	0.9622
0.0236	10.9375	1750	0.1931	0.9678	0.9790	0.9734	0.9711
0.0238	11.25	1800	0.1935	0.9610	0.9767	0.9688	0.9707
0.0206	11.5625	1850	0.1942	0.9602	0.9744	0.9672	0.9707
0.0163	11.875	1900	0.1811	0.9641	0.9790	0.9715	0.9737
0.0131	12.1875	1950	0.1731	0.9671	0.9806	0.9738	0.9758
0.017	12.5	2000	0.1835	0.9574	0.9783	0.9677	0.9707
0.0117	12.8125	2050	0.2070	0.9617	0.9759	0.9688	0.9669
0.0074	13.125	2100	0.2049	0.9617	0.9759	0.9688	0.9669
0.0093	13.4375	2150	0.2004	0.9625	0.9775	0.9700	0.9699
0.0157	13.75	2200	0.1948	0.9678	0.9790	0.9734	0.9728
0.01	14.0625	2250	0.1870	0.9648	0.9790	0.9719	0.9737
0.0047	14.375	2300	0.2000	0.9617	0.9752	0.9684	0.9699
0.0163	14.6875	2350	0.1951	0.9632	0.9767	0.9699	0.9711
0.0118	15.0	2400	0.1936	0.9640	0.9767	0.9703	0.9716
0.0093	15.3125	2450	0.1932	0.9625	0.9759	0.9692	0.9711
0.0025	15.625	2500	0.1929	0.9625	0.9759	0.9692	0.9711

Framework versions

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