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: 4.5570

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
5.2892	0.22	50	4.6213
4.4958	0.44	100	4.1766
4.1602	0.66	150	3.8565
3.9538	0.88	200	3.5805
3.5903	1.11	250	3.4594
3.1555	1.33	300	3.1305
3.0041	1.55	350	2.9394
2.8133	1.77	400	2.7611
2.5517	1.99	450	2.6359
1.9883	2.21	500	2.6822
2.1057	2.43	550	2.4430
2.0872	2.65	600	2.4388
1.7591	2.88	650	2.3726
1.5244	3.1	700	2.3325
1.5257	3.32	750	2.7668
1.3295	3.54	800	2.2800
1.3167	3.76	850	2.3108
1.2051	3.98	900	2.6429
1.0532	4.2	950	3.1483
1.0167	4.42	1000	2.2237
1.1105	4.65	1050	2.3531
1.006	4.87	1100	2.8026
0.8193	5.09	1150	3.0897
0.7502	5.31	1200	3.5147
0.9308	5.53	1250	2.5877
0.6495	5.75	1300	2.8749
0.7933	5.97	1350	2.7216
0.7199	6.19	1400	3.1257
0.5458	6.42	1450	3.5710
0.7406	6.64	1500	2.5855
0.5537	6.86	1550	2.9661
0.493	7.08	1600	3.3691
0.4938	7.3	1650	3.1937
0.5407	7.52	1700	3.3585
0.5243	7.74	1750	2.7650
0.4177	7.96	1800	3.2618
0.4094	8.19	1850	3.2418
0.4061	8.41	1900	3.2710
0.3248	8.63	1950	3.3732
0.3787	8.85	2000	3.3467
0.2709	9.07	2050	3.1261
0.2655	9.29	2100	3.4552
0.2552	9.51	2150	3.4344
0.3321	9.73	2200	3.5558
0.4165	9.96	2250	3.0775
0.2052	10.18	2300	3.4985
0.2642	10.4	2350	3.8930
0.2461	10.62	2400	3.7029
0.2895	10.84	2450	3.4143
0.1041	11.06	2500	3.6652
0.1316	11.28	2550	3.8867
0.1165	11.5	2600	4.1669
0.1925	11.73	2650	4.3376
0.371	11.95	2700	3.9125
0.1793	12.17	2750	4.0051
0.1771	12.39	2800	3.8161
0.2158	12.61	2850	3.8786
0.1838	12.83	2900	3.7687
0.1364	13.05	2950	4.1103
0.0364	13.27	3000	4.1468
0.2488	13.5	3050	4.1149
0.1046	13.72	3100	4.1887
0.1254	13.94	3150	4.0097
0.033	14.16	3200	4.3023
0.0373	14.38	3250	4.3530
0.0911	14.6	3300	4.2621
0.1014	14.82	3350	3.8815
0.0238	15.04	3400	4.4097
0.1141	15.27	3450	4.4720
0.0278	15.49	3500	4.4407
0.0673	15.71	3550	4.4176
0.0691	15.93	3600	4.4863
0.0379	16.15	3650	4.6924
0.057	16.37	3700	4.5305
0.1087	16.59	3750	4.5050
0.0047	16.81	3800	4.5652
0.051	17.04	3850	4.5482
0.0054	17.26	3900	4.5337
0.0542	17.48	3950	4.5370
0.0132	17.7	4000	4.4744
0.0698	17.92	4050	4.4535
0.0214	18.14	4100	4.4660
0.033	18.36	4150	4.4818
0.029	18.58	4200	4.5033
0.0283	18.81	4250	4.4825
0.013	19.03	4300	4.4794
0.0538	19.25	4350	4.5183
0.0171	19.47	4400	4.5373
0.0122	19.69	4450	4.5501
0.0222	19.91	4500	4.5570

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3