Edit model card

bert-small-finetuned-xglue-ner-longer50

This model is a fine-tuned version of muhtasham/bert-small-finetuned-xglue-ner-longer20 on the wnut_17 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.7236
  • Precision: 0.6182
  • Recall: 0.4222
  • F1: 0.5018
  • Accuracy: 0.9252

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: 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: 30

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
No log 1.0 425 0.5693 0.5232 0.4581 0.4885 0.9268
0.0032 2.0 850 0.6191 0.5281 0.4498 0.4858 0.9260
0.0035 3.0 1275 0.7045 0.6011 0.4055 0.4843 0.9241
0.0056 4.0 1700 0.6715 0.5571 0.4438 0.4940 0.9261
0.004 5.0 2125 0.6537 0.5645 0.4294 0.4878 0.9256
0.0063 6.0 2550 0.6646 0.5659 0.4211 0.4829 0.9255
0.0063 7.0 2975 0.6269 0.5306 0.4354 0.4783 0.9238
0.003 8.0 3400 0.7235 0.5921 0.3959 0.4746 0.9238
0.0051 9.0 3825 0.6334 0.5330 0.4450 0.4850 0.9237
0.0047 10.0 4250 0.6408 0.5893 0.4462 0.5078 0.9271
0.004 11.0 4675 0.6721 0.5840 0.4282 0.4941 0.9255
0.0051 12.0 5100 0.6853 0.5795 0.4318 0.4949 0.9258
0.0038 13.0 5525 0.6870 0.5789 0.4211 0.4875 0.9249
0.0038 14.0 5950 0.6931 0.6032 0.4091 0.4875 0.9241
0.0033 15.0 6375 0.6502 0.5965 0.4510 0.5136 0.9266
0.0032 16.0 6800 0.6941 0.6126 0.4426 0.5139 0.9267
0.0042 17.0 7225 0.6603 0.5856 0.4462 0.5064 0.9266
0.0016 18.0 7650 0.6870 0.6121 0.4474 0.5169 0.9273
0.0028 19.0 8075 0.6922 0.5906 0.4366 0.5021 0.9250
0.0023 20.0 8500 0.7096 0.6089 0.4246 0.5004 0.9250
0.0023 21.0 8925 0.6763 0.5772 0.4426 0.5010 0.9261
0.0025 22.0 9350 0.6880 0.5696 0.4258 0.4873 0.9241
0.0018 23.0 9775 0.6759 0.5836 0.4426 0.5034 0.9259
0.0017 24.0 10200 0.7044 0.6198 0.4270 0.5057 0.9262
0.0018 25.0 10625 0.6948 0.6040 0.4306 0.5028 0.9245
0.0018 26.0 11050 0.6930 0.5948 0.4354 0.5028 0.9255
0.0018 27.0 11475 0.7077 0.6048 0.4246 0.4989 0.9250
0.0023 28.0 11900 0.7127 0.6103 0.4270 0.5025 0.9252
0.0013 29.0 12325 0.7253 0.6243 0.4234 0.5046 0.9254
0.0015 30.0 12750 0.7236 0.6182 0.4222 0.5018 0.9252

Framework versions

  • Transformers 4.21.1
  • Pytorch 1.12.1+cu113
  • Datasets 2.4.0
  • Tokenizers 0.12.1
Downloads last month
0
Hosted inference API
Token Classification
Examples
Examples
This model can be loaded on the Inference API on-demand.

Dataset used to train muhtasham/bert-small-finetuned-xglue-ner-longer50

Evaluation results