distilBERT-infoExtract

This model is a fine-tuned version of distilbert-base-cased on the conll2003 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0718
• Precision: 0.9134
• Recall: 0.9369
• F1: 0.9250
• Accuracy: 0.9832

Model description

The model can identify human name, organization and location so far (no time recognition). It was trained for 5 minutes with T4 GPU on Colab.

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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.0954	1.0	1756	0.0846	0.8880	0.9194	0.9034	0.9769
0.0498	2.0	3512	0.0699	0.9057	0.9310	0.9182	0.9815
0.031	3.0	5268	0.0718	0.9134	0.9369	0.9250	0.9832

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1