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: 4.3701

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
5.3439	0.22	50	4.6661
4.4287	0.44	100	4.2125
4.0781	0.66	150	3.7145
3.8419	0.88	200	3.5926
3.633	1.11	250	3.4266
3.1492	1.33	300	3.2833
3.2018	1.55	350	3.0237
2.8891	1.77	400	2.9976
2.6695	1.99	450	3.0654
2.211	2.21	500	2.7658
2.2555	2.43	550	2.5009
2.0293	2.65	600	2.3515
1.8663	2.88	650	2.1264
1.6332	3.1	700	2.3934
1.5298	3.32	750	2.6686
1.3594	3.54	800	2.0781
1.3923	3.76	850	2.2174
1.26	3.98	900	2.5728
0.9371	4.2	950	2.7164
0.9891	4.42	1000	2.8637
1.0822	4.65	1050	2.5435
0.9355	4.87	1100	2.7256
0.9527	5.09	1150	2.8934
0.9454	5.31	1200	3.0235
0.6458	5.53	1250	3.2280
0.8697	5.75	1300	2.8636
0.9365	5.97	1350	2.8955
0.4425	6.19	1400	3.0859
0.6329	6.42	1450	3.0695
0.7564	6.64	1500	2.5050
0.5747	6.86	1550	3.2825
0.5451	7.08	1600	3.4123
0.5432	7.3	1650	3.1163
0.3738	7.52	1700	2.8969
0.5026	7.74	1750	2.8579
0.4245	7.96	1800	3.2212
0.3145	8.19	1850	3.4482
0.516	8.41	1900	2.9995
0.2816	8.63	1950	2.9903
0.3946	8.85	2000	3.3378
0.3854	9.07	2050	3.4644
0.2191	9.29	2100	3.5034
0.3854	9.51	2150	3.4320
0.2207	9.73	2200	3.6972
0.2779	9.96	2250	3.6866
0.2837	10.18	2300	3.8988
0.1613	10.4	2350	3.8722
0.1069	10.62	2400	3.9079
0.4031	10.84	2450	3.5352
0.2129	11.06	2500	3.6764
0.1166	11.28	2550	4.1964
0.1599	11.5	2600	4.2577
0.2108	11.73	2650	3.7519
0.3207	11.95	2700	3.6609
0.2291	12.17	2750	3.5265
0.1609	12.39	2800	3.8727
0.2308	12.61	2850	3.9877
0.1461	12.83	2900	4.0395
0.1353	13.05	2950	3.8678
0.0735	13.27	3000	3.9054
0.1695	13.5	3050	3.5661
0.0129	13.72	3100	3.9625
0.1691	13.94	3150	3.7996
0.066	14.16	3200	4.2540
0.0484	14.38	3250	4.0625
0.099	14.6	3300	4.4666
0.127	14.82	3350	4.0762
0.046	15.04	3400	4.1859
0.0608	15.27	3450	4.4004
0.1002	15.49	3500	4.3228
0.0394	15.71	3550	4.4576
0.0863	15.93	3600	4.4386
0.0214	16.15	3650	4.5233
0.0661	16.37	3700	4.4493
0.0448	16.59	3750	4.3361
0.0463	16.81	3800	4.4174
0.0456	17.04	3850	4.4851
0.0144	17.26	3900	4.4655
0.0268	17.48	3950	4.4417
0.0529	17.7	4000	4.3580
0.0678	17.92	4050	4.2008
0.0039	18.14	4100	4.2346
0.0481	18.36	4150	4.2652
0.0501	18.58	4200	4.2786
0.0271	18.81	4250	4.2857
0.0322	19.03	4300	4.3047
0.0187	19.25	4350	4.3691
0.0469	19.47	4400	4.3560
0.0056	19.69	4450	4.3626
0.0099	19.91	4500	4.3701

Framework versions

• Transformers 4.35.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.14.1