Edit model card

layoutlmv2-base-uncased_finetuned_docvqa

This model is a fine-tuned version of microsoft/layoutlmv2-base-uncased on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 4.6207

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
5.3769 0.22 50 4.5459
4.3879 0.44 100 4.0099
3.9828 0.66 150 3.8817
3.7387 0.88 200 3.5334
3.4702 1.11 250 3.4670
3.0784 1.33 300 3.5327
2.8014 1.55 350 2.8722
2.673 1.77 400 2.8246
2.4546 1.99 450 2.7050
2.094 2.21 500 2.7639
1.9542 2.43 550 2.3472
1.8966 2.65 600 2.4508
1.7745 2.88 650 2.2813
1.4371 3.1 700 2.5234
1.3257 3.32 750 2.4358
1.3269 3.54 800 2.3044
1.4035 3.76 850 2.3546
1.6189 3.98 900 2.0838
0.9209 4.2 950 2.3836
0.8405 4.42 1000 3.1673
0.9808 4.65 1050 2.8038
0.8978 4.87 1100 2.7652
0.8733 5.09 1150 3.0965
0.7449 5.31 1200 2.9948
0.8173 5.53 1250 2.8631
0.8322 5.75 1300 2.6144
0.7147 5.97 1350 3.2041
0.6495 6.19 1400 3.3711
0.5458 6.42 1450 3.5480
0.5624 6.64 1500 3.3366
0.5736 6.86 1550 2.9356
0.2921 7.08 1600 3.4028
0.3883 7.3 1650 3.4411
0.3614 7.52 1700 3.2267
0.4376 7.74 1750 3.2137
0.4849 7.96 1800 3.6388
0.5035 8.19 1850 4.0089
0.36 8.41 1900 3.7903
0.2238 8.63 1950 3.7131
0.276 8.85 2000 3.8541
0.2661 9.07 2050 3.5220
0.291 9.29 2100 3.9075
0.3767 9.51 2150 3.4267
0.1326 9.73 2200 3.9069
0.2365 9.96 2250 3.5183
0.2676 10.18 2300 3.4975
0.131 10.4 2350 3.8138
0.1818 10.62 2400 3.6213
0.2988 10.84 2450 3.7380
0.0511 11.06 2500 4.1506
0.1909 11.28 2550 3.7369
0.0854 11.5 2600 3.8751
0.1291 11.73 2650 3.7143
0.1896 11.95 2700 4.0645
0.0746 12.17 2750 4.0363
0.1254 12.39 2800 4.0050
0.1588 12.61 2850 4.1739
0.1622 12.83 2900 4.2698
0.047 13.05 2950 4.5366
0.1023 13.27 3000 4.3094
0.1195 13.5 3050 4.3112
0.0753 13.72 3100 4.3137
0.0545 13.94 3150 4.4992
0.0771 14.16 3200 4.4759
0.0648 14.38 3250 4.5531
0.0935 14.6 3300 4.3725
0.1032 14.82 3350 4.3321
0.0199 15.04 3400 4.3527
0.0498 15.27 3450 4.2477
0.0688 15.49 3500 4.2340
0.0271 15.71 3550 4.2386
0.0176 15.93 3600 4.4715
0.0319 16.15 3650 4.5608
0.0061 16.37 3700 4.5767
0.0043 16.59 3750 4.6581
0.071 16.81 3800 4.5622
0.0689 17.04 3850 4.5067
0.0328 17.26 3900 4.4449
0.0784 17.48 3950 4.3684
0.0387 17.7 4000 4.4261
0.0107 17.92 4050 4.5190
0.0056 18.14 4100 4.4963
0.0261 18.36 4150 4.5995
0.0061 18.58 4200 4.6121
0.0354 18.81 4250 4.5794
0.0111 19.03 4300 4.5977
0.022 19.25 4350 4.6151
0.0628 19.47 4400 4.6061
0.0186 19.69 4450 4.6180
0.0077 19.91 4500 4.6207

Framework versions

  • Transformers 4.28.1
  • Pytorch 2.0.0+cu118
  • Datasets 2.12.0
  • Tokenizers 0.13.3
Downloads last month
6
Hosted inference API
This model can be loaded on the Inference API on-demand.