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: 5.2363

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
4.0408	0.2212	50	4.0001
4.1144	0.4425	100	3.7920
3.8854	0.6637	150	3.6503
3.6048	0.8850	200	3.3228
3.1846	1.1062	250	3.6110
2.917	1.3274	300	2.9913
2.8848	1.5487	350	2.7110
2.5842	1.7699	400	2.4111
2.1162	1.9912	450	2.4839
1.8347	2.2124	500	2.7160
1.786	2.4336	550	2.5238
1.8828	2.6549	600	2.4274
1.8181	2.8761	650	2.5544
1.5656	3.0973	700	2.4362
1.4265	3.3186	750	2.9550
1.4967	3.5398	800	3.2754
1.2732	3.7611	850	3.0296
1.3162	3.9823	900	2.6941
1.0837	4.2035	950	2.9119
1.1094	4.4248	1000	3.0181
1.1846	4.6460	1050	2.6419
1.5768	4.8673	1100	4.0184
1.4084	5.0885	1150	3.1371
0.9783	5.3097	1200	2.9210
0.984	5.5310	1250	3.0042
0.7546	5.7522	1300	3.1277
0.799	5.9735	1350	3.0501
0.6629	6.1947	1400	3.2626
0.8973	6.4159	1450	3.2922
0.6816	6.6372	1500	3.0462
0.539	6.8584	1550	3.1018
0.6871	7.0796	1600	3.1925
0.4569	7.3009	1650	3.2120
0.6451	7.5221	1700	2.9812
0.5579	7.7434	1750	3.3052
0.4851	7.9646	1800	4.1491
0.5851	8.1858	1850	3.5338
0.4344	8.4071	1900	3.4542
0.5021	8.6283	1950	3.2402
0.4699	8.8496	2000	3.3066
0.4668	9.0708	2050	3.6041
0.2258	9.2920	2100	3.6862
0.4708	9.5133	2150	3.7622
0.3933	9.7345	2200	3.7370
0.3858	9.9558	2250	3.3631
0.3359	10.1770	2300	3.6203
0.2365	10.3982	2350	3.7388
0.3147	10.6195	2400	3.8653
0.3401	10.8407	2450	4.0243
0.1644	11.0619	2500	4.1857
0.142	11.2832	2550	4.3611
0.266	11.5044	2600	4.2761
0.1592	11.7257	2650	4.3012
0.1126	11.9469	2700	4.3518
0.1409	12.1681	2750	4.4466
0.0731	12.3894	2800	4.3459
0.1243	12.6106	2850	4.3446
0.2672	12.8319	2900	4.3548
0.228	13.0531	2950	4.1020
0.0622	13.2743	3000	4.4363
0.1287	13.4956	3050	4.5345
0.1974	13.7168	3100	4.6727
0.2213	13.9381	3150	4.3807
0.1551	14.1593	3200	4.4805
0.1295	14.3805	3250	4.7027
0.0664	14.6018	3300	4.7583
0.1159	14.8230	3350	4.3252
0.02	15.0442	3400	4.6594
0.0438	15.2655	3450	4.8679
0.0495	15.4867	3500	5.1235
0.1143	15.7080	3550	5.1614
0.1405	15.9292	3600	5.1302
0.0351	16.1504	3650	5.0780
0.1258	16.3717	3700	5.1000
0.0387	16.5929	3750	5.0849
0.0809	16.8142	3800	4.9809
0.0955	17.0354	3850	5.0030
0.0347	17.2566	3900	5.0040
0.0716	17.4779	3950	4.9608
0.0417	17.6991	4000	5.0922
0.1394	17.9204	4050	5.1081
0.0612	18.1416	4100	5.1859
0.0057	18.3628	4150	5.2126
0.0965	18.5841	4200	5.1589
0.0131	18.8053	4250	5.1224
0.0922	19.0265	4300	5.1521
0.0353	19.2478	4350	5.1961
0.0351	19.4690	4400	5.2249
0.0161	19.6903	4450	5.2304
0.0095	19.9115	4500	5.2363

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1