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: 4.6448

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: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
5.2021	0.2212	50	4.6324
4.5313	0.4425	100	4.1367
4.1524	0.6637	150	3.8836
3.9207	0.8850	200	3.6920
3.5517	1.1062	250	3.8563
3.295	1.3274	300	3.2045
3.0938	1.5487	350	3.1921
2.9264	1.7699	400	2.8674
2.5353	1.9912	450	2.8867
2.1442	2.2124	500	2.7470
1.8984	2.4336	550	2.4924
1.886	2.6549	600	2.6026
1.9277	2.8761	650	2.6891
1.6546	3.0973	700	2.4442
1.4824	3.3186	750	2.8104
1.3533	3.5398	800	2.6744
1.5249	3.7611	850	2.1151
1.3263	3.9823	900	2.5105
0.9831	4.2035	950	2.8372
0.9544	4.4248	1000	2.2467
1.0167	4.6460	1050	3.0175
0.9889	4.8673	1100	2.6414
0.7828	5.0885	1150	2.8631
0.7675	5.3097	1200	2.8171
0.7695	5.5310	1250	3.0892
0.5891	5.7522	1300	3.1601
0.8397	5.9735	1350	3.0463
0.6106	6.1947	1400	3.3519
0.7397	6.4159	1450	3.4455
0.5613	6.6372	1500	3.1579
0.5931	6.8584	1550	3.3272
0.6042	7.0796	1600	3.0458
0.3754	7.3009	1650	3.3260
0.349	7.5221	1700	3.3049
0.5256	7.7434	1750	3.4301
0.4256	7.9646	1800	3.4376
0.333	8.1858	1850	3.4952
0.2256	8.4071	1900	3.7613
0.3261	8.6283	1950	3.3898
0.6199	8.8496	2000	3.3443
0.4307	9.0708	2050	3.1757
0.1569	9.2920	2100	3.7978
0.4755	9.5133	2150	3.5794
0.2493	9.7345	2200	3.5829
0.2686	9.9558	2250	3.5064
0.3662	10.1770	2300	3.2991
0.2353	10.3982	2350	3.4224
0.0991	10.6195	2400	4.2513
0.407	10.8407	2450	3.5800
0.1471	11.0619	2500	3.6337
0.1352	11.2832	2550	3.9379
0.2922	11.5044	2600	3.8454
0.1113	11.7257	2650	4.1881
0.1325	11.9469	2700	4.0861
0.1598	12.1681	2750	4.1164
0.0822	12.3894	2800	4.0703
0.1181	12.6106	2850	3.9423
0.2272	12.8319	2900	4.1349
0.1706	13.0531	2950	4.0460
0.1051	13.2743	3000	4.1329
0.0349	13.4956	3050	4.2074
0.2101	13.7168	3100	4.0685
0.1001	13.9381	3150	4.3431
0.1109	14.1593	3200	4.3210
0.0264	14.3805	3250	4.5687
0.1321	14.6018	3300	4.4580
0.0979	14.8230	3350	4.5390
0.0905	15.0442	3400	4.4641
0.0706	15.2655	3450	4.5589
0.0386	15.4867	3500	4.4396
0.0411	15.7080	3550	4.4250
0.0349	15.9292	3600	4.5973
0.1208	16.1504	3650	4.5193
0.0248	16.3717	3700	4.5689
0.0175	16.5929	3750	4.6381
0.0266	16.8142	3800	4.6431
0.0236	17.0354	3850	4.6552
0.0075	17.2566	3900	4.6977
0.0514	17.4779	3950	4.6455
0.0252	17.6991	4000	4.6360
0.0507	17.9204	4050	4.6566
0.0273	18.1416	4100	4.5838
0.0612	18.3628	4150	4.5459
0.0168	18.5841	4200	4.5979
0.0384	18.8053	4250	4.6035
0.0445	19.0265	4300	4.5904
0.0403	19.2478	4350	4.6543
0.0095	19.4690	4400	4.6552
0.0212	19.6903	4450	4.6426
0.0037	19.9115	4500	4.6448

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3