layoutlmv2-base-uncased_finetuned_docvqa

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 4.6973

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: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
5.3322	0.22	50	4.6214
4.4006	0.44	100	4.1705
3.9967	0.66	150	3.8115
3.8349	0.88	200	3.5599
3.4185	1.11	250	3.2811
3.1129	1.33	300	3.3014
2.9186	1.55	350	3.2583
2.8937	1.77	400	2.6806
2.5872	1.99	450	2.4617
1.934	2.21	500	2.4098
1.9086	2.43	550	2.3926
1.7456	2.65	600	2.4048
1.6092	2.88	650	2.2064
1.5114	3.1	700	2.3882
1.3176	3.32	750	2.2211
1.2236	3.54	800	3.5792
1.4123	3.76	850	2.0391
1.2434	3.98	900	2.3563
0.8727	4.2	950	2.6344
0.9026	4.42	1000	3.1071
1.0698	4.65	1050	3.1795
0.9226	4.87	1100	3.0897
1.0639	5.09	1150	3.2822
0.7568	5.31	1200	3.0442
0.8192	5.53	1250	2.8516
0.5137	5.75	1300	3.1520
0.7053	5.97	1350	2.7512
0.4587	6.19	1400	2.8772
0.4965	6.42	1450	3.1186
0.4079	6.64	1500	3.0887
0.6135	6.86	1550	3.0537
0.7518	7.08	1600	3.2562
0.336	7.3	1650	3.3709
0.618	7.52	1700	2.5948
0.3473	7.74	1750	3.1784
0.3811	7.96	1800	3.3420
0.2749	8.19	1850	3.5958
0.2916	8.41	1900	3.7811
0.2634	8.63	1950	3.8935
0.4594	8.85	2000	3.4889
0.2957	9.07	2050	3.5717
0.1503	9.29	2100	3.8861
0.2986	9.51	2150	3.8007
0.3732	9.73	2200	3.5632
0.282	9.96	2250	3.1815
0.1781	10.18	2300	3.8386
0.1678	10.4	2350	3.9624
0.1813	10.62	2400	4.1113
0.3601	10.84	2450	3.9760
0.2374	11.06	2500	3.8321
0.1251	11.28	2550	4.0945
0.0741	11.5	2600	4.1810
0.2846	11.73	2650	3.9240
0.1964	11.95	2700	3.7833
0.1413	12.17	2750	3.9830
0.0696	12.39	2800	4.4270
0.1401	12.61	2850	4.6113
0.1267	12.83	2900	4.3713
0.0689	13.05	2950	4.4582
0.1636	13.27	3000	4.6710
0.1935	13.5	3050	4.2833
0.0936	13.72	3100	4.5357
0.1001	13.94	3150	4.2660
0.0509	14.16	3200	4.3863
0.0193	14.38	3250	4.5554
0.031	14.6	3300	4.4455
0.0163	14.82	3350	4.4985
0.1558	15.04	3400	4.5935
0.0707	15.27	3450	4.4791
0.0089	15.49	3500	4.6358
0.0881	15.71	3550	4.5644
0.1348	15.93	3600	4.5336
0.0176	16.15	3650	4.5465
0.0679	16.37	3700	4.4872
0.0138	16.59	3750	4.3819
0.0211	16.81	3800	4.5024
0.0506	17.04	3850	4.4761
0.0051	17.26	3900	4.5898
0.075	17.48	3950	4.5427
0.0089	17.7	4000	4.6244
0.0298	17.92	4050	4.6440
0.0034	18.14	4100	4.6486
0.0594	18.36	4150	4.6438
0.0179	18.58	4200	4.6552
0.0081	18.81	4250	4.6508
0.0046	19.03	4300	4.6808
0.0618	19.25	4350	4.6856
0.0033	19.47	4400	4.6763
0.0062	19.69	4450	4.6892
0.0404	19.91	4500	4.6973

Framework versions

• Transformers 4.30.2
• Pytorch 1.10.0+cu113
• Datasets 2.13.2
• Tokenizers 0.13.3