deberta-finetuned-ner

baptiste/deberta-finetuned-ner

deberta-finetuned-ner is a token classification model from baptiste. Use it when you need labels on individual words, such as names. It is set up for transformers. The card lists the license as mit.

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

Downloads · 30 days

1% of all-time downloads

All-time downloads

3.6K

Public

Parameters

139M

2.2 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin555 MB · 50%

Parameter types

How the weights are stored.

F32139M · 100%

Try a prompt

Datasets

conll2003

Model card

From the Hugging Face model README

deberta-finetuned-ner

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

Loss: 0.0515
Precision: 0.9577
Recall: 0.9652
F1: 0.9614
Accuracy: 0.9907

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
0.0742	1.0	1756	0.0526	0.9390	0.9510	0.9450	0.9868
0.0374	2.0	3512	0.0528	0.9421	0.9554	0.9487	0.9879
0.0205	3.0	5268	0.0505	0.9505	0.9636	0.9570	0.9900
0.0089	4.0	7024	0.0528	0.9531	0.9636	0.9583	0.9898
0.0076	5.0	8780	0.0515	0.9577	0.9652	0.9614	0.9907

Framework versions

Transformers 4.20.1
Pytorch 1.12.0+cu113
Datasets 2.3.2
Tokenizers 0.12.1

Task: Token Classification
Library: transformers
Type: deberta
License: mit
Created: Jul 16, 2022
Updated: Sep 11, 2023