Open-source Weather Classification Model weather_classification_ViT - Superb at Image Classification with Nearly 97% Accuracy!

Weather Classification ViT

Developed by methane6923

A weather classification model fine-tuned based on Google's ViT model, excelling in image classification tasks with an accuracy of 96.79%.

Image Classification

Transformers

Open Source License:Apache-2.0 #High-precision weather classification #ViT image recognition #Fine-tuned pre-trained model

Downloads 66

Release Time : 7/8/2024

Model Overview

This model is a fine-tuned version of Google's ViT-base-patch16-224-in21k model on a weather image dataset, specifically designed for weather condition image classification tasks.

Model Features

High accuracy

Achieves an outstanding accuracy of 96.79% on the evaluation set.

Based on ViT architecture

Utilizes the Vision Transformer architecture, which has powerful image feature extraction capabilities.

Efficient training

Requires only 4 training epochs to achieve high performance, demonstrating high training efficiency.

Model Capabilities

Weather image classification

Multi-category image recognition

Use Cases

Meteorological monitoring

Automatic weather recognition

Used for automatically identifying and classifying weather conditions such as sunny, rainy, cloudy, etc.

Accuracy: 96.79%

Smart agriculture

Farmland weather monitoring

Monitors weather conditions in farmland areas to provide decision-making support for agricultural production.

🚀 weather_classification_ViT

This model is a fine - tuned version of [google/vit - base - patch16 - 224 - in21k](https://huggingface.co/google/vit - base - patch16 - 224 - in21k) on the imagefolder dataset. It can effectively classify weather images, achieving high accuracy in evaluation.

🚀 Quick Start

This model is a fine - tuned version of [google/vit - base - patch16 - 224 - in21k](https://huggingface.co/google/vit - base - patch16 - 224 - in21k) on the imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 0.1268
Accuracy: 0.9679
Precision: 0.9679
Recall: 0.9679
F1: 0.9679
Auc: 0.9974

📚 Documentation

Model Information

Property	Details
Model Type	Fine - tuned version of google/vit - base - patch16 - 224 - in21k
Training Data	imagefolder dataset
Metrics	Accuracy, Precision, Recall, F1, Auc
License	Apache - 2.0

Training Procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0002
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
num_epochs: 4
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Auc
0.2811	0.2288	100	0.3139	0.8958	0.9147	0.8958	0.8970	0.9903
0.1396	0.4577	200	0.2454	0.9278	0.9307	0.9278	0.9282	0.9919
0.3761	0.6865	300	0.2952	0.9072	0.9117	0.9072	0.9071	0.9889
0.2365	0.9153	400	0.1797	0.9444	0.9447	0.9444	0.9445	0.9940
0.2528	1.1442	500	0.2470	0.9278	0.9307	0.9278	0.9278	0.9924
0.2364	1.3730	600	0.2448	0.9261	0.9306	0.9261	0.9264	0.9934
0.34	1.6018	700	0.1986	0.9404	0.9409	0.9404	0.9405	0.9929
0.2001	1.8307	800	0.1525	0.9542	0.9548	0.9542	0.9539	0.9960
0.0958	2.0595	900	0.1783	0.9507	0.9515	0.9507	0.9505	0.9952
0.1862	2.2883	1000	0.1654	0.9553	0.9558	0.9553	0.9551	0.9952
0.1021	2.5172	1100	0.1654	0.9462	0.9472	0.9462	0.9459	0.9958
0.1178	2.7460	1200	0.1591	0.9525	0.9536	0.9525	0.9523	0.9960
0.0474	2.9748	1300	0.1299	0.9633	0.9635	0.9633	0.9633	0.9975
0.046	3.2037	1400	0.1384	0.9628	0.9628	0.9628	0.9627	0.9972
0.0294	3.4325	1500	0.1388	0.9645	0.9644	0.9645	0.9644	0.9969
0.1833	3.6613	1600	0.1346	0.9633	0.9634	0.9633	0.9633	0.9971
0.0548	3.8902	1700	0.1268	0.9679	0.9679	0.9679	0.9679	0.9974

Framework versions

Transformers 4.41.2
Pytorch 2.3.0+cu121
Datasets 2.20.0
Tokenizers 0.19.1

📄 License

This project is licensed under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご