bert-finetuned-ner2

NineArtsDragon/bert-finetuned-ner2

bert-finetuned-ner2 is a token classification model from NineArtsDragon. Use it when you need labels on individual words, such as names. It is set up for transformers. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

2% of all-time downloads

All-time downloads

368

Public

Repo size

1.7 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin431 MB · 100%

At a glance

Task: Token Classification
Library: transformers
License: apache-2.0
Model type: bert
Access: Public
Created: Nov 11, 2022
Updated: Nov 11, 2022
SHA: 640f017c

Try a prompt

Model card

From the Hugging Face model README

bert-finetuned-ner2

This model is a fine-tuned version of bert-base-cased on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.0015
Precision: 0.9709
Recall: 0.9904
F1: 0.9806
Accuracy: 0.9995

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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	108	0.0035	0.9565	0.9876	0.9718	0.9990
No log	2.0	216	0.0019	0.9628	0.9904	0.9764	0.9994
No log	3.0	324	0.0016	0.9764	0.9904	0.9834	0.9995
No log	4.0	432	0.0017	0.9736	0.9866	0.9801	0.9995
0.0166	5.0	540	0.0015	0.9709	0.9904	0.9806	0.9995

Framework versions

Transformers 4.24.0
Pytorch 1.12.1+cu113
Datasets 2.6.1
Tokenizers 0.13.2

Task: Token Classification
Library: transformers
Type: bert
License: apache-2.0
Created: Nov 11, 2022
Updated: Nov 11, 2022