README.md

metadata

library_name: transformers
license: cc-by-nc-sa-4.0
base_model: microsoft/layoutlmv3-large
tags:
  - generated_from_trainer
datasets:
  - mp-02/cord
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: layoutlmv3-large-cord2
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: mp-02/cord
          type: mp-02/cord
        metrics:
          - name: Precision
            type: precision
            value: 0.9810074318744839
          - name: Recall
            type: recall
            value: 0.9842584921292461
          - name: F1
            type: f1
            value: 0.9826302729528537
          - name: Accuracy
            type: accuracy
            value: 0.9817017383348582

layoutlmv3-large-cord2

This model is a fine-tuned version of microsoft/layoutlmv3-large on the mp-02/cord dataset. It achieves the following results on the evaluation set:

• Loss: 0.1205
• Precision: 0.9810
• Recall: 0.9843
• F1: 0.9826
• Accuracy: 0.9817

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: 1e-05
• train_batch_size: 10
• eval_batch_size: 10
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 3000

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.25	100	0.4870	0.8359	0.8691	0.8522	0.8518
No log	2.5	200	0.1731	0.9505	0.9702	0.9602	0.9584
No log	3.75	300	0.1432	0.9559	0.9693	0.9626	0.9684
No log	5.0	400	0.0925	0.9745	0.9809	0.9777	0.9808
0.4385	6.25	500	0.1295	0.9695	0.9760	0.9727	0.9748
0.4385	7.5	600	0.1169	0.9696	0.9785	0.9740	0.9758
0.4385	8.75	700	0.1040	0.9769	0.9826	0.9798	0.9812
0.4385	10.0	800	0.1268	0.9696	0.9785	0.9740	0.9771
0.4385	11.25	900	0.1514	0.9687	0.9735	0.9711	0.9716
0.0431	12.5	1000	0.1230	0.9794	0.9843	0.9818	0.9812
0.0431	13.75	1100	0.1327	0.9786	0.9834	0.9810	0.9794
0.0431	15.0	1200	0.1300	0.9761	0.9809	0.9785	0.9794
0.0431	16.25	1300	0.1312	0.9802	0.9843	0.9822	0.9812
0.0431	17.5	1400	0.1358	0.9761	0.9818	0.9789	0.9799
0.0146	18.75	1500	0.1205	0.9810	0.9843	0.9826	0.9817
0.0146	20.0	1600	0.1481	0.9753	0.9826	0.9790	0.9785
0.0146	21.25	1700	0.1710	0.9728	0.9768	0.9748	0.9726
0.0146	22.5	1800	0.1969	0.9622	0.9693	0.9657	0.9680
0.0146	23.75	1900	0.1613	0.9745	0.9801	0.9773	0.9780
0.0084	25.0	2000	0.1713	0.9720	0.9793	0.9757	0.9758
0.0084	26.25	2100	0.1414	0.9761	0.9826	0.9794	0.9794
0.0084	27.5	2200	0.1510	0.9737	0.9809	0.9773	0.9780
0.0084	28.75	2300	0.1435	0.9794	0.9851	0.9822	0.9803
0.0084	30.0	2400	0.1685	0.9728	0.9793	0.9761	0.9758
0.0047	31.25	2500	0.1620	0.9728	0.9793	0.9761	0.9762
0.0047	32.5	2600	0.1549	0.9761	0.9818	0.9789	0.9780
0.0047	33.75	2700	0.1566	0.9777	0.9826	0.9802	0.9785
0.0047	35.0	2800	0.1627	0.9769	0.9826	0.9798	0.9785
0.0047	36.25	2900	0.1580	0.9777	0.9826	0.9802	0.9785
0.0034	37.5	3000	0.1592	0.9777	0.9826	0.9802	0.9785

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu118
• Datasets 2.21.0
• Tokenizers 0.19.1