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: 4.7623

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
5.2477	0.2212	50	4.4955
4.4322	0.4425	100	4.0859
4.2098	0.6637	150	3.8177
3.8664	0.8850	200	3.5806
3.5251	1.1062	250	3.5032
3.1732	1.3274	300	3.2472
3.0841	1.5487	350	3.0947
2.8536	1.7699	400	2.7800
2.4276	1.9912	450	2.7769
2.06	2.2124	500	2.6578
1.8173	2.4336	550	2.6715
2.1107	2.6549	600	2.5620
2.1352	2.8761	650	2.3209
1.5368	3.0973	700	2.2305
1.3107	3.3186	750	2.6482
1.4519	3.5398	800	2.3794
1.2756	3.7611	850	2.3672
1.2282	3.9823	900	2.2342
1.0882	4.2035	950	2.7453
0.9957	4.4248	1000	2.7899
1.0055	4.6460	1050	2.7979
0.9377	4.8673	1100	2.5045
1.0285	5.0885	1150	2.4579
0.6299	5.3097	1200	2.7204
0.8789	5.5310	1250	2.6098
0.5642	5.7522	1300	2.7831
0.6949	5.9735	1350	3.0893
0.4063	6.1947	1400	2.8284
0.5117	6.4159	1450	3.0634
0.4416	6.6372	1500	3.3999
0.4999	6.8584	1550	3.2898
0.5086	7.0796	1600	3.4221
0.3996	7.3009	1650	3.0418
0.235	7.5221	1700	3.3613
0.4907	7.7434	1750	3.1062
0.3033	7.9646	1800	4.0306
0.3901	8.1858	1850	3.8258
0.3625	8.4071	1900	3.2560
0.3074	8.6283	1950	3.6874
0.3582	8.8496	2000	3.2337
0.2091	9.0708	2050	3.2660
0.2416	9.2920	2100	3.4408
0.1241	9.5133	2150	3.6883
0.2945	9.7345	2200	3.5552
0.2575	9.9558	2250	3.3925
0.258	10.1770	2300	3.8662
0.1662	10.3982	2350	3.6742
0.1491	10.6195	2400	4.3579
0.2379	10.8407	2450	4.1496
0.0899	11.0619	2500	4.2631
0.026	11.2832	2550	4.3676
0.1356	11.5044	2600	4.1160
0.0734	11.7257	2650	3.8254
0.2507	11.9469	2700	3.9717
0.1241	12.1681	2750	3.7671
0.0207	12.3894	2800	3.9668
0.0662	12.6106	2850	4.0811
0.1262	12.8319	2900	3.9894
0.0483	13.0531	2950	4.0627
0.0889	13.2743	3000	4.1365
0.0311	13.4956	3050	4.1390
0.0992	13.7168	3100	4.0020
0.1021	13.9381	3150	3.8962
0.109	14.1593	3200	4.2122
0.0164	14.3805	3250	4.3584
0.0663	14.6018	3300	4.1452
0.0702	14.8230	3350	4.2793
0.0435	15.0442	3400	4.3782
0.0504	15.2655	3450	4.3851
0.0185	15.4867	3500	4.6016
0.0795	15.7080	3550	4.5381
0.049	15.9292	3600	4.2093
0.0608	16.1504	3650	4.3391
0.0953	16.3717	3700	4.2657
0.0603	16.5929	3750	4.4624
0.0312	16.8142	3800	4.3063
0.0038	17.0354	3850	4.4603
0.0271	17.2566	3900	4.4354
0.0094	17.4779	3950	4.6563
0.019	17.6991	4000	4.7925
0.045	17.9204	4050	4.6123
0.0112	18.1416	4100	4.6376
0.0348	18.3628	4150	4.6756
0.0216	18.5841	4200	4.7026
0.009	18.8053	4250	4.7217
0.0356	19.0265	4300	4.7260
0.0479	19.2478	4350	4.7143
0.0114	19.4690	4400	4.7547
0.0069	19.6903	4450	4.7605
0.0053	19.9115	4500	4.7623

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.0+cu121
• Datasets 3.0.2
• Tokenizers 0.19.1