layoutlmv2-base-uncased_finetuned_docvqa

This model is a fine-tuned version of microsoft/layoutlmv2-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 5.2024

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
5.2725	0.22	50	4.6295
4.532	0.44	100	4.2254
4.0834	0.66	150	3.8568
3.9787	0.88	200	3.6523
3.6224	1.11	250	3.8143
3.2397	1.33	300	3.1685
3.0931	1.55	350	3.0822
2.9593	1.77	400	2.9521
2.5904	1.99	450	2.7995
2.1358	2.21	500	2.4827
2.0137	2.43	550	2.3651
1.951	2.65	600	2.2448
1.684	2.88	650	2.3455
1.6708	3.1	700	2.5003
1.3927	3.32	750	2.3116
1.5237	3.54	800	2.6236
1.2826	3.76	850	2.4859
1.5274	3.98	900	2.1857
1.0727	4.2	950	2.5041
0.9465	4.42	1000	2.7958
1.0889	4.65	1050	2.3797
0.9121	4.87	1100	2.7570
0.8847	5.09	1150	3.0968
0.8864	5.31	1200	2.8488
0.8693	5.53	1250	2.6848
0.5451	5.75	1300	3.3272
0.8121	5.97	1350	3.6097
0.6214	6.19	1400	3.1954
0.5576	6.42	1450	3.2427
0.5576	6.64	1500	3.4471
0.4858	6.86	1550	3.2469
0.5947	7.08	1600	3.2522
0.4889	7.3	1650	3.3459
0.3126	7.52	1700	3.9616
0.3291	7.74	1750	3.9943
0.5337	7.96	1800	3.6498
0.2384	8.19	1850	4.2966
0.3566	8.41	1900	4.1365
0.3539	8.63	1950	4.1291
0.3219	8.85	2000	4.3024
0.2307	9.07	2050	4.1780
0.1922	9.29	2100	4.4078
0.1721	9.51	2150	4.2569
0.1541	9.73	2200	4.2138
0.3044	9.96	2250	4.2793
0.2642	10.18	2300	4.2676
0.1725	10.4	2350	4.0887
0.2223	10.62	2400	3.9813
0.2107	10.84	2450	4.1089
0.3262	11.06	2500	3.9025
0.383	11.28	2550	4.1511
0.1437	11.5	2600	4.1774
0.2707	11.73	2650	4.1749
0.1692	11.95	2700	4.3012
0.1651	12.17	2750	4.3723
0.1388	12.39	2800	4.3273
0.1072	12.61	2850	4.7238
0.1748	12.83	2900	4.3425
0.1053	13.05	2950	4.0696
0.1929	13.27	3000	4.6322
0.028	13.5	3050	4.4843
0.0207	13.72	3100	4.9324
0.0662	13.94	3150	4.9421
0.0644	14.16	3200	4.8991
0.0321	14.38	3250	4.7757
0.0567	14.6	3300	4.9158
0.0552	14.82	3350	5.0722
0.0695	15.04	3400	5.0160
0.081	15.27	3450	5.1969
0.084	15.49	3500	5.2285
0.0372	15.71	3550	5.2621
0.1193	15.93	3600	5.1806
0.0166	16.15	3650	5.2799
0.029	16.37	3700	5.2543
0.048	16.59	3750	5.1176
0.143	16.81	3800	5.1800
0.033	17.04	3850	5.1635
0.0424	17.26	3900	5.1982
0.004	17.48	3950	5.2322
0.0143	17.7	4000	5.2242
0.0261	17.92	4050	5.3110
0.0076	18.14	4100	5.3329
0.0036	18.36	4150	5.3355
0.0182	18.58	4200	5.3223
0.0466	18.81	4250	5.2396
0.0036	19.03	4300	5.2409
0.0278	19.25	4350	5.2128
0.015	19.47	4400	5.2227
0.0394	19.69	4450	5.2018
0.0034	19.91	4500	5.2024

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0