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: 4.9512

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
5.2951	0.2212	50	4.7074
4.5271	0.4425	100	4.1032
4.1303	0.6637	150	3.9369
3.8633	0.8850	200	3.5651
3.4747	1.1062	250	3.6489
3.4294	1.3274	300	3.1484
3.1319	1.5487	350	2.9161
2.8258	1.7699	400	2.8137
2.6171	1.9912	450	2.9083
2.4222	2.2124	500	3.0206
2.1271	2.4336	550	2.6056
1.9083	2.6549	600	2.3721
1.9685	2.8761	650	2.3992
1.6307	3.0973	700	2.7301
1.4481	3.3186	750	2.3406
1.4984	3.5398	800	2.6880
1.7125	3.7611	850	2.4609
1.3863	3.9823	900	2.3533
1.2462	4.2035	950	2.5440
1.0761	4.4248	1000	2.7272
1.1144	4.6460	1050	1.9797
0.9787	4.8673	1100	2.4466
0.8171	5.0885	1150	2.9085
0.8736	5.3097	1200	2.9959
0.9349	5.5310	1250	2.3175
0.7439	5.7522	1300	3.0023
0.7007	5.9735	1350	3.1554
0.4858	6.1947	1400	3.3196
0.7693	6.4159	1450	3.3589
0.4525	6.6372	1500	3.1226
0.442	6.8584	1550	3.5817
0.5321	7.0796	1600	3.4303
0.4044	7.3009	1650	3.5700
0.3062	7.5221	1700	4.0335
0.5116	7.7434	1750	3.6335
0.3481	7.9646	1800	3.5987
0.2781	8.1858	1850	4.0488
0.3054	8.4071	1900	3.4694
0.3572	8.6283	1950	3.7605
0.3524	8.8496	2000	3.9977
0.3551	9.0708	2050	3.7106
0.1455	9.2920	2100	3.8854
0.3655	9.5133	2150	3.6476
0.348	9.7345	2200	3.5709
0.299	9.9558	2250	3.6914
0.1781	10.1770	2300	3.7616
0.1812	10.3982	2350	3.7328
0.3197	10.6195	2400	4.3628
0.2139	10.8407	2450	4.3765
0.0266	11.0619	2500	4.6882
0.0659	11.2832	2550	4.5060
0.2312	11.5044	2600	4.1473
0.1366	11.7257	2650	4.2983
0.2264	11.9469	2700	4.2400
0.165	12.1681	2750	4.0799
0.0443	12.3894	2800	4.2081
0.178	12.6106	2850	4.0656
0.2517	12.8319	2900	4.2621
0.1208	13.0531	2950	4.6150
0.0767	13.2743	3000	4.6056
0.0789	13.4956	3050	4.2637
0.0492	13.7168	3100	4.3896
0.1782	13.9381	3150	4.3427
0.0968	14.1593	3200	4.5666
0.0287	14.3805	3250	4.6159
0.0516	14.6018	3300	4.4566
0.1008	14.8230	3350	4.5481
0.052	15.0442	3400	4.6010
0.0334	15.2655	3450	4.6442
0.027	15.4867	3500	4.7174
0.0791	15.7080	3550	4.5374
0.147	15.9292	3600	4.6980
0.1186	16.1504	3650	4.5936
0.0373	16.3717	3700	4.6215
0.0445	16.5929	3750	4.4737
0.0325	16.8142	3800	4.6176
0.0842	17.0354	3850	4.7082
0.0097	17.2566	3900	4.7821
0.0707	17.4779	3950	4.7606
0.0583	17.6991	4000	4.8275
0.0222	17.9204	4050	4.8638
0.0413	18.1416	4100	4.8912
0.0097	18.3628	4150	4.9439
0.0075	18.5841	4200	4.9960
0.0735	18.8053	4250	4.9482
0.0821	19.0265	4300	5.0201
0.0254	19.2478	4350	4.9772
0.1105	19.4690	4400	4.9629
0.0649	19.6903	4450	4.9474
0.025	19.9115	4500	4.9512

Framework versions

• Transformers 4.45.2
• Pytorch 2.5.0
• Datasets 3.0.1
• Tokenizers 0.20.1