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: 4.4147

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
5.2445	0.2212	50	4.5117
4.4073	0.4425	100	4.2219
4.1282	0.6637	150	3.8024
3.8337	0.8850	200	3.6136
3.5425	1.1062	250	3.4874
3.3022	1.3274	300	3.1532
3.2627	1.5487	350	3.0497
2.9534	1.7699	400	3.0351
2.7062	1.9912	450	3.0241
2.3171	2.2124	500	2.6357
2.0499	2.4336	550	2.3957
1.9222	2.6549	600	2.4198
1.8248	2.8761	650	2.5430
1.5168	3.0973	700	4.2506
1.4417	3.3186	750	2.3309
1.3993	3.5398	800	2.3422
1.3945	3.7611	850	2.0276
1.138	3.9823	900	2.3541
1.2168	4.2035	950	2.8071
1.1358	4.4248	1000	2.6772
1.0205	4.6460	1050	2.7978
0.9784	4.8673	1100	3.3150
1.001	5.0885	1150	2.5239
0.8487	5.3097	1200	2.9815
1.0126	5.5310	1250	3.2436
0.6363	5.7522	1300	3.2784
0.9224	5.9735	1350	3.4480
0.6946	6.1947	1400	3.1487
0.6052	6.4159	1450	3.4397
0.5203	6.6372	1500	2.9999
0.6589	6.8584	1550	3.2889
0.6399	7.0796	1600	3.1920
0.4313	7.3009	1650	2.9790
0.3867	7.5221	1700	3.4399
0.5132	7.7434	1750	3.0626
0.4955	7.9646	1800	3.2692
0.3658	8.1858	1850	3.4662
0.2021	8.4071	1900	3.7119
0.394	8.6283	1950	3.5633
0.4442	8.8496	2000	3.7246
0.3807	9.0708	2050	3.5174
0.2692	9.2920	2100	3.8268
0.3595	9.5133	2150	3.6366
0.3495	9.7345	2200	3.5126
0.3814	9.9558	2250	3.4845
0.2319	10.1770	2300	3.5154
0.1587	10.3982	2350	3.9049
0.2771	10.6195	2400	3.9095
0.2156	10.8407	2450	3.9481
0.1906	11.0619	2500	3.9076
0.2064	11.2832	2550	3.9890
0.1756	11.5044	2600	3.8731
0.1934	11.7257	2650	3.8914
0.1177	11.9469	2700	4.0169
0.2135	12.1681	2750	3.6795
0.1198	12.3894	2800	3.9709
0.1219	12.6106	2850	3.7425
0.1073	12.8319	2900	4.2397
0.1428	13.0531	2950	3.9107
0.0728	13.2743	3000	4.2249
0.0516	13.4956	3050	3.9716
0.1044	13.7168	3100	4.2036
0.2026	13.9381	3150	3.8552
0.1182	14.1593	3200	4.0365
0.0368	14.3805	3250	4.3629
0.0331	14.6018	3300	4.4697
0.1629	14.8230	3350	3.9966
0.0619	15.0442	3400	4.1223
0.0167	15.2655	3450	4.2150
0.0602	15.4867	3500	4.1427
0.1045	15.7080	3550	3.9883
0.0629	15.9292	3600	4.1485
0.0492	16.1504	3650	3.9531
0.0657	16.3717	3700	4.2826
0.0354	16.5929	3750	4.1867
0.0327	16.8142	3800	4.1699
0.0045	17.0354	3850	4.1904
0.0139	17.2566	3900	4.2937
0.0373	17.4779	3950	4.1179
0.039	17.6991	4000	4.1837
0.0717	17.9204	4050	4.2483
0.0316	18.1416	4100	4.2423
0.0041	18.3628	4150	4.2651
0.0268	18.5841	4200	4.3379
0.0156	18.8053	4250	4.3978
0.0265	19.0265	4300	4.3942
0.0461	19.2478	4350	4.4056
0.0195	19.4690	4400	4.4082
0.0105	19.6903	4450	4.4160
0.0387	19.9115	4500	4.4147

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1