Segformer-b4-wall Open-source Visual Segmentation Model - Accurately and Efficiently Achieve Wall Image Segmentation Tasks

Segformer B4 Wall

Developed by leftattention

A vision segmentation model based on the SegFormer architecture, specifically designed for wall image segmentation tasks, achieving high accuracy and IoU metrics on the evaluation set.

Image Segmentation

Transformers

#High-precision image segmentation #Wall recognition #Real-time semantic segmentation

Downloads 18

Release Time : 6/25/2024

Model Overview

This model is a vision segmentation model trained on the SegFormer-B4 architecture, primarily used for wall image segmentation tasks. The model performs excellently on the evaluation set, with an average accuracy of 94.48% and an average IoU of 89.93%.

Model Features

High-precision segmentation

Achieves an average accuracy of 94.48% and an average IoU of 89.93% on the evaluation set, demonstrating excellent segmentation performance.

Efficient training

Uses the Adam optimizer and linear learning rate scheduler to achieve rapid convergence within 50 training epochs.

Stable performance

Stable validation metrics during training indicate the model has strong generalization capabilities.

Model Capabilities

Wall image segmentation

Pixel-level classification

Visual scene understanding

Use Cases

Construction industry

Wall condition assessment

Automatically segments wall regions for building quality inspection and maintenance.

Accurately identifies wall regions, providing a foundation for subsequent analysis.

Interior design

Wall decoration planning

Precisely segments wall regions to assist in interior decoration design and planning.

🚀 segformer-b4-wall

This is a model trained from scratch for image segmentation. It achieves excellent results on the evaluation set.

🚀 Quick Start

This model, segformer-b4-wall, was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.1537
Mean Accuracy: 0.9448
Mean Iou: 0.8993
Overall Accuracy: 0.9558
Per Category Accuracy: [0.9648476610683054, 0.9680509025433003, 0.9015647356112896, nan]
Per Category Iou: [0.9294668192886654, 0.9344825387850888, 0.8340281823830938, nan]

📚 Documentation

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

🔧 Technical Details

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 6e-05
train_batch_size: 64
eval_batch_size: 64
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Accuracy	Mean Iou	Overall Accuracy	Per Category Accuracy	Per Category Iou
0.1398	5.3476	1000	0.1477	0.9424	0.8733	0.9420	[0.9375947027923643, 0.962438818648652, 0.9270677962243152, nan]	[0.9071928258269675, 0.9154732958813474, 0.7971633247503161, nan]
0.1114	10.6952	2000	0.1329	0.9426	0.8878	0.9498	[0.9551513266050631, 0.9606741248023447, 0.9120448217426163, nan]	[0.9197608920879746, 0.9255854097692368, 0.818153830444766, nan]
0.0683	16.0428	3000	0.1353	0.9473	0.8921	0.9516	[0.9527839457434386, 0.9691455504455139, 0.9198476394516605, nan]	[0.922537499674425, 0.926305870761282, 0.8273726843249476, nan]
0.0753	21.3904	4000	0.1311	0.9437	0.8959	0.9540	[0.9633835386385788, 0.9611760655179852, 0.9066569940696604, nan]	[0.9267602358926313, 0.9312805978213234, 0.8297698871401628, nan]
0.0505	26.7380	5000	0.1397	0.9442	0.8971	0.9545	[0.9627544499461427, 0.967327419780526, 0.9024453947068249, nan]	[0.9272910775593762, 0.9304849186604474, 0.8333807013974415, nan]
0.0427	32.0856	6000	0.1414	0.9455	0.8992	0.9555	[0.9640187847053339, 0.9652081246861538, 0.9074073950598316, nan]	[0.9289147168722637, 0.9321577805497577, 0.8366507705917902, nan]
0.0556	37.4332	7000	0.1477	0.9452	0.8984	0.9552	[0.9629165900233977, 0.9697602413261539, 0.9029026554269718, nan]	[0.9285106797857617, 0.9331322728249959, 0.833620894806762, nan]
0.0424	42.7807	8000	0.1484	0.9439	0.8990	0.9557	[0.9653151526182964, 0.96949089540134, 0.8967977175922358, nan]	[0.9292691886525306, 0.9343666443212755, 0.83323737535253, nan]
0.053	48.1283	9000	0.1537	0.9448	0.8993	0.9558	[0.9648476610683054, 0.9680509025433003, 0.9015647356112896, nan]	[0.9294668192886654, 0.9344825387850888, 0.8340281823830938, nan]

Framework versions

Transformers 4.40.2
Pytorch 2.3.0+cu121
Datasets 2.17.0
Tokenizers 0.19.1

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