deberta-v3

NaveenTNS/deberta-v3

deberta-v3 is a text classification model from NaveenTNS. Use it when you need a label for a piece of text. 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

20% of all-time downloads

All-time downloads

Public

Parameters

184M

19.2 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors738 MB · 99%

At a glance

Task: Text Classification
Library: transformers
License: mit
Model type: deberta-v2
Access: Public
Created: Nov 27, 2025
Updated: Nov 28, 2025
SHA: 4cd16a9c

Try a prompt

Base models

microsoft/deberta-v3-base

Model card

From the Hugging Face model README

deberta-v3

This model is a fine-tuned version of microsoft/deberta-v3-base on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.7985
Accuracy: 0.8491
Precision: 0.8425
Recall: 0.8585
F1 Macro: 0.8492
F1 Weighted: 0.8500

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: 16
eval_batch_size: 16
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 5
mixed_precision_training: Native AMP
label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1 Macro	F1 Weighted
1.1036	0.1964	500	1.0877	0.6719	0.3725	0.4317	0.3861	0.6390
0.9225	0.3928	1000	0.9070	0.7776	0.7052	0.6301	0.6187	0.7624
0.8686	0.5892	1500	0.8580	0.7976	0.7383	0.6755	0.6912	0.7892
0.8721	0.7855	2000	0.8183	0.8081	0.7776	0.7662	0.7587	0.8096
0.8199	0.9819	2500	0.8037	0.8197	0.7848	0.7828	0.7795	0.8191
0.7514	1.1783	3000	0.8130	0.8286	0.7946	0.8180	0.8044	0.8291
0.7619	1.3747	3500	0.7802	0.8365	0.8160	0.8205	0.8169	0.8355
0.7497	1.5711	4000	0.7689	0.8388	0.8431	0.8166	0.8278	0.8381
0.7306	1.7675	4500	0.7749	0.8410	0.8410	0.8239	0.8311	0.8411
0.7348	1.9639	5000	0.7516	0.8381	0.8231	0.8314	0.8256	0.8386
0.6903	2.1603	5500	0.7617	0.8397	0.8327	0.8361	0.8322	0.8409
0.6861	2.3566	6000	0.7734	0.8423	0.8309	0.8467	0.8368	0.8434
0.6655	2.5530	6500	0.7673	0.8429	0.8416	0.8434	0.8380	0.8444
0.6825	2.7494	7000	0.7797	0.8447	0.8220	0.8554	0.8371	0.8459
0.6598	2.9458	7500	0.7666	0.8479	0.8491	0.8372	0.8397	0.8480
0.6259	3.1422	8000	0.7779	0.8441	0.8530	0.8297	0.8385	0.8437
0.608	3.3386	8500	0.7985	0.8491	0.8425	0.8585	0.8492	0.8500
0.6134	3.5350	9000	0.7903	0.8469	0.8494	0.8428	0.8441	0.8476
0.6165	3.7313	9500	0.8059	0.8455	0.8393	0.8514	0.8428	0.8465
0.6112	3.9277	10000	0.8120	0.8391	0.8476	0.8273	0.8354	0.8372
0.5856	4.1241	10500	0.8042	0.8458	0.8535	0.8419	0.8465	0.8454
0.5853	4.3205	11000	0.8283	0.8485	0.8473	0.8483	0.8454	0.8492
0.5906	4.5169	11500	0.8140	0.8482	0.8478	0.8490	0.8479	0.8478
0.5643	4.7133	12000	0.8120	0.8480	0.8452	0.8503	0.8472	0.8485
0.5844	4.9097	12500	0.8156	0.8486	0.8457	0.8503	0.8474	0.8488

Framework versions

Transformers 4.46.1
Pytorch 2.3.0+cu121
Datasets 4.4.1
Tokenizers 0.20.3

Task: Text Classification
Library: transformers
Type: deberta-v2
License: mit
Created: Nov 27, 2025
Updated: Nov 28, 2025