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: 2.9710

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: 3

Training results

Training Loss	Epoch	Step	Validation Loss
6.12	0.0221	5	5.8130
5.8642	0.0442	10	5.5440
5.739	0.0664	15	5.3407
5.24	0.0885	20	5.1918
5.2382	0.1106	25	5.0621
5.0044	0.1327	30	4.9099
4.8735	0.1549	35	4.7621
4.5752	0.1770	40	4.7436
4.9789	0.1991	45	4.6436
5.3167	0.2212	50	4.5981
5.1172	0.2434	55	4.6847
4.7205	0.2655	60	4.5649
4.5686	0.2876	65	4.5079
4.774	0.3097	70	4.3704
4.2153	0.3319	75	4.3057
4.5881	0.3540	80	4.2297
4.4437	0.3761	85	4.2064
4.1528	0.3982	90	4.1870
4.2176	0.4204	95	4.2060
4.145	0.4425	100	4.1738
4.487	0.4646	105	4.1157
4.215	0.4867	110	4.1209
4.2117	0.5088	115	4.0113
4.2441	0.5310	120	3.9862
3.8206	0.5531	125	4.0846
4.418	0.5752	130	3.9696
3.8883	0.5973	135	3.9478
3.9334	0.6195	140	3.9126
4.2097	0.6416	145	3.8813
4.0268	0.6637	150	3.9252
4.126	0.6858	155	3.8643
4.0452	0.7080	160	3.9387
3.9409	0.7301	165	3.8127
3.9958	0.7522	170	3.7989
3.8162	0.7743	175	3.8034
3.5596	0.7965	180	3.8704
4.081	0.8186	185	3.7822
4.1374	0.8407	190	3.7431
4.1355	0.8628	195	3.7494
4.0031	0.8850	200	3.7118
4.0624	0.9071	205	3.8061
3.7152	0.9292	210	3.7471
4.301	0.9513	215	3.9199
4.0595	0.9735	220	3.7722
4.1836	0.9956	225	3.6203
3.6276	1.0177	230	3.6073
3.4787	1.0398	235	3.5770
3.3633	1.0619	240	3.5469
3.2999	1.0841	245	3.6939
3.4353	1.1062	250	3.7339
3.663	1.1283	255	3.5301
3.283	1.1504	260	3.5172
3.5445	1.1726	265	3.5076
3.1999	1.1947	270	3.5342
3.4036	1.2168	275	3.4955
3.31	1.2389	280	3.4295
3.3661	1.2611	285	3.4398
3.2727	1.2832	290	3.4223
3.3522	1.3053	295	3.4298
3.1652	1.3274	300	3.4076
2.9084	1.3496	305	3.3806
3.2943	1.3717	310	3.3692
3.2965	1.3938	315	3.3601
3.2069	1.4159	320	3.3893
3.285	1.4381	325	3.4980
3.1824	1.4602	330	3.4643
3.4277	1.4823	335	3.3506
3.1088	1.5044	340	3.2569
3.1225	1.5265	345	3.2182
2.9275	1.5487	350	3.3265
3.0438	1.5708	355	3.3541
3.2014	1.5929	360	3.2822
3.0306	1.6150	365	3.2362
2.9716	1.6372	370	3.2018
3.0015	1.6593	375	3.1488
2.8433	1.6814	380	3.1138
3.0251	1.7035	385	3.0836
3.0188	1.7257	390	3.1137
2.8269	1.7478	395	3.1072
3.2609	1.7699	400	3.1077
2.8849	1.7920	405	3.1659
2.6843	1.8142	410	3.2268
2.9859	1.8363	415	3.2020
2.5574	1.8584	420	3.1025
2.9709	1.8805	425	3.1188
3.1064	1.9027	430	3.0549
2.7347	1.9248	435	2.9965
2.6075	1.9469	440	2.9799
2.9998	1.9690	445	3.0093
2.4259	1.9912	450	3.1338
2.5547	2.0133	455	3.3225
2.9147	2.0354	460	3.3662
3.004	2.0575	465	3.2570
2.4481	2.0796	470	3.1761
2.5156	2.1018	475	3.1332
2.5695	2.1239	480	3.0219
2.3243	2.1460	485	3.0122
2.4268	2.1681	490	3.0692
2.3157	2.1903	495	3.1625
2.6856	2.2124	500	3.1868
2.3567	2.2345	505	3.1789
2.3799	2.2566	510	3.1141
2.3814	2.2788	515	3.0845
2.6517	2.3009	520	3.0001
2.8808	2.3230	525	2.9786
2.2501	2.3451	530	3.0351
2.4319	2.3673	535	3.0998
2.4569	2.3894	540	3.1180
2.1893	2.4115	545	3.0840
2.5029	2.4336	550	3.0379
2.5414	2.4558	555	2.9775
2.414	2.4779	560	2.9478
2.4732	2.5	565	2.9530
2.7319	2.5221	570	2.9462
2.3984	2.5442	575	2.9199
2.1631	2.5664	580	2.9257
2.1815	2.5885	585	2.9564
2.4294	2.6106	590	2.9570
2.298	2.6327	595	2.9290
2.2535	2.6549	600	2.9287
2.1774	2.6770	605	2.9196
2.2014	2.6991	610	2.9162
2.1422	2.7212	615	2.9466
2.494	2.7434	620	2.9844
2.6516	2.7655	625	2.9899
2.1923	2.7876	630	2.9580
2.3944	2.8097	635	2.9432
2.0892	2.8319	640	2.9422
2.129	2.8540	645	2.9597
2.4273	2.8761	650	2.9647
2.1467	2.8982	655	2.9614
2.1653	2.9204	660	2.9596
2.1992	2.9425	665	2.9642
2.1921	2.9646	670	2.9702
2.2585	2.9867	675	2.9710

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1