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: 4.9097

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
5.3187	0.22	50	4.6843
4.5031	0.44	100	4.1274
4.128	0.66	150	3.7921
3.9091	0.88	200	3.6115
3.6889	1.11	250	3.5696
3.1707	1.33	300	3.1811
3.0354	1.55	350	2.8562
2.7856	1.77	400	2.7444
2.4529	1.99	450	2.7386
1.9655	2.21	500	2.6058
1.9712	2.43	550	2.3241
1.9305	2.65	600	2.2414
1.7701	2.88	650	2.1765
1.3844	3.1	700	2.7470
1.4493	3.32	750	2.3821
1.2487	3.54	800	2.0868
1.3189	3.76	850	2.5860
1.3709	3.98	900	2.4280
1.1087	4.2	950	2.5985
0.9481	4.42	1000	3.3342
1.0106	4.65	1050	3.0318
0.9703	4.87	1100	2.7671
0.788	5.09	1150	3.0565
0.6952	5.31	1200	3.3631
0.8017	5.53	1250	3.2544
0.6489	5.75	1300	3.0772
0.7419	5.97	1350	2.6226
0.558	6.19	1400	3.5186
0.5727	6.42	1450	3.3600
0.5322	6.64	1500	3.5181
0.6472	6.86	1550	3.6967
0.5805	7.08	1600	3.2425
0.4877	7.3	1650	3.4871
0.273	7.52	1700	3.5272
0.5527	7.74	1750	3.0758
0.2936	7.96	1800	3.3492
0.2585	8.19	1850	3.4836
0.5038	8.41	1900	3.3159
0.3943	8.63	1950	3.3977
0.3129	8.85	2000	3.8042
0.2977	9.07	2050	3.7062
0.2695	9.29	2100	3.7420
0.363	9.51	2150	3.6655
0.1834	9.73	2200	3.7858
0.3347	9.96	2250	3.9257
0.2159	10.18	2300	3.9075
0.2691	10.4	2350	3.8001
0.2965	10.62	2400	3.6177
0.3078	10.84	2450	3.8440
0.1397	11.06	2500	4.0490
0.052	11.28	2550	4.2137
0.1741	11.5	2600	4.2273
0.2956	11.73	2650	3.8075
0.2098	11.95	2700	4.1653
0.1466	12.17	2750	4.2080
0.1378	12.39	2800	4.0473
0.1864	12.61	2850	4.0665
0.1938	12.83	2900	4.1019
0.2332	13.05	2950	3.9249
0.0486	13.27	3000	4.2374
0.223	13.5	3050	4.1405
0.1387	13.72	3100	4.1833
0.1447	13.94	3150	4.0949
0.0265	14.16	3200	4.2652
0.0253	14.38	3250	4.5294
0.1781	14.6	3300	4.5310
0.0676	14.82	3350	4.3897
0.0692	15.04	3400	4.3176
0.1093	15.27	3450	4.3505
0.0219	15.49	3500	4.4436
0.0834	15.71	3550	4.4736
0.0728	15.93	3600	4.5234
0.0612	16.15	3650	4.6432
0.0844	16.37	3700	4.5729
0.0269	16.59	3750	4.7230
0.037	16.81	3800	4.7512
0.0532	17.04	3850	4.8247
0.0042	17.26	3900	4.8263
0.0684	17.48	3950	4.7495
0.0058	17.7	4000	4.7748
0.0158	17.92	4050	4.8827
0.0059	18.14	4100	4.8746
0.0217	18.36	4150	4.8981
0.0725	18.58	4200	4.9908
0.0399	18.81	4250	4.8712
0.0478	19.03	4300	4.9006
0.0212	19.25	4350	4.9160
0.0269	19.47	4400	4.9076
0.0317	19.69	4450	4.9081
0.0233	19.91	4500	4.9097

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1