metadata

license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model_index:
  - name: chinese-address-ner
    results:
      - task:
          name: Token Classification
          type: token-classification
        metric:
          name: Accuracy
          type: accuracy
          value: 0.9852459016393442

chinese-address-ner

This model is a fine-tuned version of hfl/chinese-roberta-wwm-ext on an unkown dataset. It achieves the following results on the evaluation set:

Loss: 0.0999
Precision: 0.9739
Recall: 0.9849
F1: 0.9794
Accuracy: 0.9852

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: 2e-05
train_batch_size: 50
eval_batch_size: 50
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.0656	0.14	1	0.1061	0.9665	0.9811	0.9738	0.9844
0.1305	0.29	2	0.1096	0.9630	0.9811	0.9720	0.9836
0.1009	0.43	3	0.0999	0.9739	0.9849	0.9794	0.9852
0.0844	0.57	4	0.0911	0.9739	0.9849	0.9794	0.9852
0.0773	0.71	5	0.0858	0.9703	0.9849	0.9775	0.9852
0.0997	0.86	6	0.0815	0.9739	0.9849	0.9794	0.9861
0.0904	1.0	7	0.0795	0.9739	0.9849	0.9794	0.9861

Framework versions

Transformers 4.8.2
Pytorch 1.7.0
Datasets 1.9.0
Tokenizers 0.10.3