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: 5.1691

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
5.2266	0.22	50	4.7202
4.4028	0.44	100	4.1366
4.0483	0.66	150	3.8471
3.7553	0.88	200	3.5072
3.4046	1.11	250	3.5053
3.1543	1.33	300	3.1483
3.0577	1.55	350	3.0745
2.8076	1.77	400	2.8592
2.456	1.99	450	2.7019
1.9787	2.21	500	2.7553
1.9308	2.43	550	2.3374
1.7637	2.65	600	2.2109
1.7719	2.88	650	2.9137
1.6661	3.1	700	2.7946
1.3154	3.32	750	2.5750
1.3015	3.54	800	2.4883
1.2136	3.76	850	2.1466
1.253	3.98	900	2.2343
0.96	4.2	950	2.6481
0.9083	4.42	1000	2.2891
0.9441	4.65	1050	2.8459
0.9041	4.87	1100	3.0106
0.8727	5.09	1150	2.7765
0.6496	5.31	1200	3.0633
0.7388	5.53	1250	2.7464
0.5012	5.75	1300	3.3843
0.6762	5.97	1350	3.6035
0.4907	6.19	1400	3.4269
0.5893	6.42	1450	3.2352
0.4987	6.64	1500	3.2802
0.3867	6.86	1550	3.8191
0.6091	7.08	1600	3.4476
0.4088	7.3	1650	3.5099
0.4135	7.52	1700	3.4519
0.3859	7.74	1750	3.4147
0.335	7.96	1800	3.8082
0.2068	8.19	1850	4.3927
0.3149	8.41	1900	3.7065
0.2526	8.63	1950	3.6056
0.451	8.85	2000	3.7065
0.3792	9.07	2050	3.8738
0.2299	9.29	2100	3.8282
0.3064	9.51	2150	3.6586
0.26	9.73	2200	3.9155
0.3218	9.96	2250	3.5863
0.2826	10.18	2300	3.5095
0.149	10.4	2350	3.4537
0.1213	10.62	2400	3.8778
0.2157	10.84	2450	3.8106
0.2149	11.06	2500	4.2672
0.1212	11.28	2550	4.2534
0.1664	11.5	2600	4.3033
0.1487	11.73	2650	3.9483
0.1088	11.95	2700	3.9682
0.0791	12.17	2750	4.2143
0.0734	12.39	2800	4.2175
0.128	12.61	2850	4.2613
0.1851	12.83	2900	3.9094
0.1215	13.05	2950	4.2045
0.0438	13.27	3000	4.5802
0.0107	13.5	3050	4.7988
0.0606	13.72	3100	5.0228
0.0819	13.94	3150	4.9309
0.1375	14.16	3200	4.8995
0.0729	14.38	3250	4.7900
0.0832	14.6	3300	4.6417
0.0687	14.82	3350	4.8781
0.0516	15.04	3400	4.9231
0.056	15.27	3450	5.1059
0.0672	15.49	3500	5.0563
0.1036	15.71	3550	4.5889
0.1551	15.93	3600	4.4488
0.0509	16.15	3650	4.8964
0.0505	16.37	3700	4.9259
0.0295	16.59	3750	4.8617
0.0766	16.81	3800	4.7973
0.0103	17.04	3850	4.9249
0.0148	17.26	3900	4.9319
0.0105	17.48	3950	5.1000
0.0341	17.7	4000	4.9627
0.0547	17.92	4050	5.0149
0.0106	18.14	4100	5.0924
0.0045	18.36	4150	5.1550
0.0072	18.58	4200	5.1620
0.0428	18.81	4250	5.1546
0.0443	19.03	4300	5.1506
0.0244	19.25	4350	5.1720
0.0096	19.47	4400	5.1640
0.0265	19.69	4450	5.1647
0.0057	19.91	4500	5.1691

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2