Tengisbold's picture
update model card README.md
b03af5f
metadata
language:
  - mn
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: roberta-base-ner-demo
    results: []

roberta-base-ner-demo

This model is a fine-tuned version of bayartsogt/mongolian-roberta-base on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1349
  • Precision: 0.9210
  • Recall: 0.9330
  • F1: 0.9269
  • Accuracy: 0.9788

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: 16
  • eval_batch_size: 32
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
0.1615 1.0 477 0.0917 0.8327 0.8817 0.8565 0.9680
0.0615 2.0 954 0.0917 0.8432 0.8918 0.8668 0.9701
0.0404 3.0 1431 0.0961 0.8411 0.8970 0.8682 0.9710
0.0227 4.0 1908 0.1056 0.9081 0.9262 0.9171 0.9770
0.0133 5.0 2385 0.1117 0.8781 0.9108 0.8942 0.9744
0.0083 6.0 2862 0.1231 0.9111 0.9288 0.9199 0.9775
0.0061 7.0 3339 0.1263 0.9110 0.9293 0.9200 0.9776
0.0071 8.0 3816 0.1316 0.9197 0.9302 0.9249 0.9783
0.0031 9.0 4293 0.1335 0.9228 0.9327 0.9277 0.9790
0.0021 10.0 4770 0.1349 0.9210 0.9330 0.9269 0.9788

Framework versions

  • Transformers 4.27.4
  • Pytorch 2.0.0+cu118
  • Datasets 2.11.0
  • Tokenizers 0.13.3