README.md

metadata

tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: layoutlmv1-er-ner
    results: []

layoutlmv1-er-ner

This model is a fine-tuned version of renjithks/layoutlmv1-cord-ner on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2092
• Precision: 0.7202
• Recall: 0.7238
• F1: 0.7220
• Accuracy: 0.9639

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: 8
• 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	Precision	Recall	F1	Accuracy
No log	1.0	41	0.2444	0.4045	0.3996	0.4020	0.9226
No log	2.0	82	0.1640	0.5319	0.6098	0.5682	0.9455
No log	3.0	123	0.1531	0.6324	0.6614	0.6466	0.9578
No log	4.0	164	0.1440	0.6927	0.6743	0.6834	0.9620
No log	5.0	205	0.1520	0.6750	0.6958	0.6853	0.9613
No log	6.0	246	0.1597	0.6840	0.6987	0.6913	0.9605
No log	7.0	287	0.1910	0.7002	0.6887	0.6944	0.9605
No log	8.0	328	0.1860	0.6834	0.6923	0.6878	0.9609
No log	9.0	369	0.1665	0.6785	0.7102	0.6940	0.9624
No log	10.0	410	0.1816	0.7016	0.7052	0.7034	0.9624
No log	11.0	451	0.1808	0.6913	0.7166	0.7038	0.9638
No log	12.0	492	0.2165	0.712	0.7023	0.7071	0.9628
0.1014	13.0	533	0.2135	0.6979	0.7109	0.7043	0.9613
0.1014	14.0	574	0.2154	0.6906	0.7109	0.7006	0.9612
0.1014	15.0	615	0.2118	0.6902	0.7016	0.6958	0.9615
0.1014	16.0	656	0.2091	0.6985	0.7080	0.7032	0.9623
0.1014	17.0	697	0.2104	0.7118	0.7123	0.7121	0.9630
0.1014	18.0	738	0.2081	0.7129	0.7231	0.7179	0.9638
0.1014	19.0	779	0.2093	0.7205	0.7231	0.7218	0.9638
0.1014	20.0	820	0.2092	0.7202	0.7238	0.7220	0.9639

Framework versions

• Transformers 4.18.0
• Pytorch 1.11.0
• Datasets 2.1.0
• Tokenizers 0.12.1