Modelpublic

pooh/layoutlmv2-base-uncased_finetuned_docvqa

sourceHugging Facecc-by-nc-sa-4.0updated 3y agoView on Hugging Face

0likes6downloads

Model Card

layoutlmv2-base-uncasedfinetuneddocvqa

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: 4.9529

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
—trainbatchsize: 4
—evalbatchsize: 8
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
5.2903	0.22	50	4.6096
4.441	0.44	100	4.1809
4.1512	0.66	150	3.8270
3.9297	0.88	200	3.6180
3.7006	1.11	250	3.3508
3.1238	1.33	300	3.4886
3.177	1.55	350	3.0878
2.8817	1.77	400	2.8975
2.6113	1.99	450	3.1366
2.9929	2.21	500	4.2811
2.8507	2.43	550	3.1442
2.6294	2.65	600	2.7537
2.9134	2.88	650	4.0845
2.7527	3.1	700	2.6888
2.2184	3.32	750	2.6068
1.9832	3.54	800	2.3920
1.8607	3.76	850	2.3026
1.6756	3.98	900	2.4535
1.594	4.2	950	2.3539
1.3695	4.42	1000	2.9487
1.4473	4.65	1050	2.3269
1.0998	4.87	1100	2.5812
1.043	5.09	1150	2.7785
1.0655	5.31	1200	3.1658
1.2366	5.53	1250	3.5025
1.1033	5.75	1300	3.0308
1.1406	5.97	1350	2.4193
0.7332	6.19	1400	3.0098
0.7752	6.42	1450	3.0226
0.9816	6.64	1500	3.1292
0.794	6.86	1550	3.4569
0.6923	7.08	1600	3.5805
0.4034	7.3	1650	3.9237
0.4836	7.52	1700	3.4433
0.6216	7.74	1750	3.1084
0.6027	7.96	1800	3.5491
0.4783	8.19	1850	3.7448
0.4513	8.41	1900	3.4646
0.4544	8.63	1950	3.7954
0.5161	8.85	2000	3.7831
0.1872	9.07	2050	3.6736
0.506	9.29	2100	3.7390
0.2257	9.51	2150	3.9423
0.2648	9.73	2200	3.7982
0.3953	9.96	2250	3.2984
0.1601	10.18	2300	3.6460
0.2689	10.4	2350	3.9842
0.2762	10.62	2400	3.2707
0.3091	10.84	2450	3.4759
0.2036	11.06	2500	3.7818
0.1104	11.28	2550	3.8338
0.1555	11.5	2600	3.7824
0.2794	11.73	2650	3.7954
0.2728	11.95	2700	3.5966
0.2168	12.17	2750	4.2583
0.1133	12.39	2800	4.3897
0.293	12.61	2850	3.9776
0.1307	12.83	2900	4.4287
0.2012	13.05	2950	4.0434
0.1583	13.27	3000	3.8509
0.1016	13.5	3050	3.9090
0.0329	13.72	3100	4.2917
0.1034	13.94	3150	4.3789
0.0928	14.16	3200	4.4046
0.1318	14.38	3250	4.2611
0.1015	14.6	3300	4.4932
0.1499	14.82	3350	4.4150
0.1858	15.04	3400	4.1948
0.1402	15.27	3450	4.3734
0.0584	15.49	3500	4.3949
0.0288	15.71	3550	4.6144
0.0554	15.93	3600	4.8472
0.0853	16.15	3650	4.7406
0.0111	16.37	3700	5.0774
0.1094	16.59	3750	4.9672
0.0102	16.81	3800	4.9885
0.0884	17.04	3850	5.0612
0.0318	17.26	3900	5.1363
0.1083	17.48	3950	4.7403
0.0891	17.7	4000	4.6907
0.0495	17.92	4050	4.7827
0.015	18.14	4100	5.0118
0.0554	18.36	4150	4.9823
0.084	18.58	4200	4.9539
0.0714	18.81	4250	4.8877
0.0573	19.03	4300	4.9120
0.012	19.25	4350	4.9568
0.0381	19.47	4400	4.9459
0.0126	19.69	4450	4.9544
0.0591	19.91	4500	4.9529

Framework versions

—Transformers 4.32.1
—Pytorch 2.1.0
—Datasets 2.14.6
—Tokenizers 0.13.3