deberta-v3-base-ner-B

alexeyak/deberta-v3-base-ner-B

deberta-v3-base-ner-B is a token classification model from alexeyak. 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

17% of all-time downloads

All-time downloads

Public

Parameters

184M

5.2 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors735 MB · 99%

At a glance

Task: Token Classification
Library: transformers
License: mit
Model type: deberta-v2
Access: Public
Created: Dec 4, 2023
Updated: Dec 6, 2023
SHA: 0954a5a5

Try a prompt

Base models

microsoft/deberta-v3-base

Model card

From the Hugging Face model README

deberta-v3-base-ner-B

This model is a fine-tuned version of microsoft/deberta-v3-base on English part of Babelscape/multinerd dataset. It achieves the following results on the evaluation set:

Loss: 0.0294
Precision: 0.9660
Recall: 0.9751
F1: 0.9705
Accuracy: 0.9929

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: 5e-05
train_batch_size: 16
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 128
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 3
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.0336	0.13	257	0.0345	0.9245	0.9386	0.9315	0.9885
0.0309	0.25	514	0.0296	0.9161	0.9624	0.9387	0.9892
0.0281	0.38	771	0.0251	0.9316	0.9539	0.9426	0.9908
0.0221	0.5	1028	0.0259	0.9381	0.9588	0.9483	0.9910
0.0234	0.63	1285	0.0260	0.9318	0.9640	0.9477	0.9904
0.0177	0.75	1542	0.0248	0.9331	0.9665	0.9495	0.9909
0.0213	0.88	1799	0.0228	0.9522	0.9593	0.9557	0.9918
0.0252	1.0	2056	0.0233	0.9517	0.9568	0.9542	0.9917
0.0143	1.13	2313	0.0256	0.9491	0.9641	0.9565	0.9918
0.0132	1.25	2570	0.0247	0.9536	0.9627	0.9581	0.9921
0.015	1.38	2827	0.0243	0.9467	0.9640	0.9553	0.9917
0.0148	1.5	3084	0.0254	0.9475	0.9677	0.9575	0.9918
0.0143	1.63	3341	0.0252	0.9491	0.9667	0.9578	0.9920
0.0112	1.75	3598	0.0244	0.9546	0.9626	0.9586	0.9923
0.0074	1.88	3855	0.0268	0.9490	0.9680	0.9584	0.9921
0.0068	2.0	4112	0.0257	0.9577	0.9610	0.9594	0.9923
0.0079	2.13	4369	0.0296	0.9457	0.9698	0.9576	0.9919
0.0067	2.26	4626	0.0290	0.9520	0.9686	0.9602	0.9922
0.0067	2.38	4883	0.0282	0.9553	0.9653	0.9603	0.9923
0.0044	2.51	5140	0.0303	0.9600	0.9622	0.9611	0.9926
0.005	2.63	5397	0.0318	0.9488	0.9703	0.9594	0.9920
0.006	2.76	5654	0.0295	0.9564	0.9663	0.9613	0.9925
0.0059	2.88	5911	0.0304	0.9586	0.9657	0.9621	0.9925

Framework versions

Transformers 4.35.2
Pytorch 2.1.0+cu118
Datasets 2.15.0
Tokenizers 0.15.0

Task: Token Classification
Library: transformers
Type: deberta-v2
License: mit
Created: Dec 4, 2023
Updated: Dec 6, 2023