layoutlmv2-base-uncased_finetuned_docvqa

This model is a fine-tuned version of microsoft/layoutlmv2-base-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 5.0085

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
5.3352	0.22	50	4.5120
4.3566	0.44	100	4.0171
3.9989	0.66	150	3.9234
3.8014	0.88	200	3.5051
3.5509	1.11	250	3.5408
3.1372	1.33	300	3.2247
2.9307	1.55	350	3.1225
2.928	1.77	400	2.9461
2.7004	1.99	450	2.5206
2.1271	2.21	500	2.6079
2.1387	2.43	550	2.8524
1.9593	2.65	600	2.8749
2.0105	2.88	650	2.6666
1.84	3.1	700	3.0599
1.9359	3.32	750	3.0472
1.547	3.54	800	2.2308
1.4161	3.76	850	2.2889
2.1804	3.98	900	2.1462
1.0261	4.2	950	2.9056
1.392	4.42	1000	3.0021
1.3816	4.65	1050	2.6913
1.0117	4.87	1100	2.8484
1.0094	5.09	1150	2.6936
0.7316	5.31	1200	2.9901
0.9172	5.53	1250	2.6366
0.8608	5.75	1300	2.8584
0.7116	5.97	1350	3.1944
0.321	6.19	1400	3.4703
0.6663	6.42	1450	3.0456
0.6319	6.64	1500	3.3318
0.7001	6.86	1550	3.1439
0.5952	7.08	1600	3.3220
0.39	7.3	1650	3.8266
0.434	7.52	1700	3.8287
0.7599	7.74	1750	3.4079
0.52	7.96	1800	3.3982
0.5257	8.19	1850	3.5208
0.4304	8.41	1900	3.8404
0.4213	8.63	1950	3.9974
0.3033	8.85	2000	3.9492
0.2947	9.07	2050	3.9279
0.2285	9.29	2100	3.5652
0.3472	9.51	2150	3.5741
0.2644	9.73	2200	3.8685
0.3667	9.96	2250	3.5242
0.1528	10.18	2300	3.5848
0.1489	10.4	2350	3.8603
0.1984	10.62	2400	3.6773
0.3131	10.84	2450	3.7021
0.1866	11.06	2500	3.8918
0.1908	11.28	2550	3.9479
0.1955	11.5	2600	3.9596
0.1382	11.73	2650	4.1168
0.2528	11.95	2700	4.1007
0.0538	12.17	2750	4.2003
0.1354	12.39	2800	4.3118
0.1218	12.61	2850	4.1494
0.1956	12.83	2900	4.1475
0.0691	13.05	2950	4.4141
0.0526	13.27	3000	4.7115
0.0984	13.5	3050	4.6013
0.1828	13.72	3100	4.2457
0.0906	13.94	3150	4.4969
0.025	14.16	3200	4.6981
0.0149	14.38	3250	4.8642
0.123	14.6	3300	4.5326
0.0876	14.82	3350	4.5953
0.0771	15.04	3400	4.4175
0.066	15.27	3450	4.6324
0.0542	15.49	3500	4.5058
0.0293	15.71	3550	4.7244
0.0428	15.93	3600	4.9415
0.009	16.15	3650	4.9592
0.0715	16.37	3700	4.9211
0.0044	16.59	3750	4.9854
0.0767	16.81	3800	4.7985
0.0356	17.04	3850	4.7618
0.0562	17.26	3900	4.9239
0.0085	17.48	3950	4.9837
0.0114	17.7	4000	5.0808
0.0057	17.92	4050	5.0377
0.0306	18.14	4100	5.0137
0.0426	18.36	4150	4.9367
0.0429	18.58	4200	5.0050
0.0081	18.81	4250	4.9806
0.0168	19.03	4300	4.9902
0.0074	19.25	4350	4.9939
0.0075	19.47	4400	4.9986
0.0307	19.69	4450	5.0095
0.02	19.91	4500	5.0085

Framework versions

• Transformers 4.29.1
• Pytorch 1.12.1
• Datasets 2.11.0
• Tokenizers 0.11.0