SegFormer-B0 Finetuned Deprem Satellite Open-Source Model - Empowering Semantic Segmentation Analysis of Satellite Images in Earthquake Scenarios

Segformer B0 Finetuned Deprem Satellite

Developed by sawthiha

A semantic segmentation model fine-tuned on satellite image datasets based on NVIDIA SegFormer-B0 architecture, specifically designed for earthquake-related scenario analysis

Image Segmentation

Transformers

Open Source License:Other #Satellite Image Segmentation #High-Precision Semantic Segmentation #Specialized for Disaster Assessment

Downloads 21

Release Time : 1/1/2024

Model Overview

This model is an optimized deep learning model for satellite image semantic segmentation tasks, particularly suitable for earthquake disaster assessment scenarios. It can accurately identify and segment different land cover categories in satellite images.

Model Features

High-Precision Segmentation

Achieves 98.49% mean Intersection over Union (mIoU) and 99.33% overall accuracy on the evaluation dataset

Efficient Inference

Processes 10.988 samples per second during evaluation, suitable for real-time applications

Domain Adaptation

Specifically fine-tuned on the deprem_satellite_semantic_whu_dataset, optimized for earthquake scenario analysis

Model Capabilities

Satellite Image Analysis

Semantic Segmentation

Disaster Assessment

Land Cover Identification

Use Cases

Disaster Response

Earthquake Damage Assessment

Analyze damage to buildings and infrastructure after earthquakes through satellite images

Can accurately segment damaged areas to support rescue decision-making

Geographic Information Systems

Land Use Classification

Automatically classify and label different land types in satellite images

🚀 segformer-b0-finetuned-deprem-satellite

This model is a fine - tuned version of a pre - trained model, which aims to solve the problem of image segmentation in the deprem - related satellite image scenarios. It provides high - accuracy segmentation results on the specific dataset, bringing value to relevant vision tasks.

🚀 Quick Start

This model is a fine - tuned version of [nvidia/segformer - b0 - finetuned - ade - 512 - 512](https://huggingface.co/nvidia/segformer - b0 - finetuned - ade - 512 - 512) on the deprem - ml/deprem_satellite_semantic_whu_dataset dataset. It achieves the following results on the evaluation set:

Property	Details
eval_loss	0.0641
eval_mean_iou	0.9849
eval_mean_accuracy	0.9933
eval_overall_accuracy	0.9933
eval_runtime	94.2835
eval_samples_per_second	10.988
eval_steps_per_second	2.206
epoch	4.18
step	1980

✨ Features

Fine - tuned Model: Based on the pre - trained nvidia/segformer - b0 - finetuned - ade - 512 - 512 model, it is further optimized for the deprem - ml/deprem_satellite_semantic_whu_dataset.
High - accuracy Results: Achieves excellent performance in evaluation metrics such as mean IoU and accuracy.

📦 Installation

No installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

No code examples are provided in the original document, so this section is skipped.

📚 Documentation

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

🔧 Technical Details

Training hyperparameters

The following hyperparameters were used during training:

Property	Details
learning_rate	7e - 05
train_batch_size	10
eval_batch_size	5
seed	42
optimizer	Adam with betas=(0.9, 0.999) and epsilon = 1e - 08
lr_scheduler_type	linear
num_epochs	5

Framework versions

Property	Details
Transformers	4.36.2
Pytorch	2.1.2
Datasets	2.16.1
Tokenizers	0.15.0

📄 License

The license for this model is other.

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