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: 4.4423

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
5.2693	0.22	50	4.4222
4.3703	0.44	100	4.1079
4.1363	0.66	150	3.9209
3.7332	0.88	200	3.6332
3.4591	1.11	250	3.5577
3.1781	1.33	300	3.1319
3.3388	1.55	350	3.0866
2.8356	1.77	400	2.7820
2.4286	1.99	450	2.8378
2.0496	2.21	500	2.5224
1.9469	2.43	550	2.5281
1.8342	2.65	600	2.5674
1.6589	2.88	650	2.2914
1.6939	3.1	700	2.4427
1.3883	3.32	750	2.5626
1.3944	3.54	800	2.3736
1.2459	3.76	850	2.7994
1.5218	3.98	900	2.5390
1.1471	4.2	950	2.5951
0.8888	4.42	1000	2.7430
0.971	4.65	1050	2.5219
1.0425	4.87	1100	2.5474
0.7665	5.09	1150	2.9321
0.8039	5.31	1200	2.7369
0.6426	5.53	1250	3.1309
0.6628	5.75	1300	3.1167
0.906	5.97	1350	3.8550
0.6223	6.19	1400	3.4892
0.6274	6.42	1450	3.2927
0.4732	6.64	1500	3.4192
0.5962	6.86	1550	3.2867
0.6761	7.08	1600	3.0610
0.4096	7.3	1650	3.5926
0.457	7.52	1700	3.2824
0.3721	7.74	1750	3.4383
0.4547	7.96	1800	3.4794
0.4231	8.19	1850	3.7591
0.3292	8.41	1900	3.8104
0.4401	8.63	1950	3.7450
0.446	8.85	2000	3.5815
0.3362	9.07	2050	3.6245
0.1832	9.29	2100	3.7162
0.2085	9.51	2150	3.8565
0.3248	9.73	2200	3.4577
0.4722	9.96	2250	3.6518
0.2575	10.18	2300	3.8701
0.2336	10.4	2350	3.7511
0.2864	10.62	2400	3.7999
0.2091	10.84	2450	3.8716
0.2371	11.06	2500	3.7909
0.1582	11.28	2550	4.0463
0.2519	11.5	2600	3.9798
0.1223	11.73	2650	4.3331
0.1838	11.95	2700	4.1601
0.1204	12.17	2750	4.2846
0.1797	12.39	2800	4.1595
0.123	12.61	2850	4.2625
0.2177	12.83	2900	4.0050
0.1728	13.05	2950	4.0885
0.1525	13.27	3000	3.9733
0.0388	13.5	3050	4.1072
0.0788	13.72	3100	4.2446
0.1629	13.94	3150	4.0483
0.0377	14.16	3200	4.2435
0.0966	14.38	3250	4.1510
0.0943	14.6	3300	4.2591
0.048	14.82	3350	4.1876
0.097	15.04	3400	4.2489
0.0188	15.27	3450	4.3612
0.1163	15.49	3500	4.2931
0.0754	15.71	3550	4.3306
0.1044	15.93	3600	4.2243
0.0316	16.15	3650	4.3932
0.005	16.37	3700	4.4173
0.0389	16.59	3750	4.3939
0.0505	16.81	3800	4.3207
0.0501	17.04	3850	4.3601
0.0491	17.26	3900	4.3211
0.0048	17.48	3950	4.3425
0.0043	17.7	4000	4.3461
0.0309	17.92	4050	4.3733
0.0246	18.14	4100	4.3912
0.0055	18.36	4150	4.4020
0.0078	18.58	4200	4.4256
0.0057	18.81	4250	4.4462
0.0352	19.03	4300	4.4558
0.0451	19.25	4350	4.4557
0.063	19.47	4400	4.4395
0.0123	19.69	4450	4.4428
0.0291	19.91	4500	4.4423

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.1+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0