S

Segformer B3 1024x1024 City 160k

Developed by smp-hub
A semantic segmentation model based on the Segformer architecture, optimized for the Cityscapes dataset
Downloads 14
Release Time : 11/29/2024

Model Overview

This model is a semantic segmentation model based on the Transformer architecture, specifically designed for pixel-level classification tasks in urban street scene images. It adopts Segformer's hybrid design, combining the advantages of convolution and Transformer.

Model Features

Hybrid Architecture Design
Combines the advantages of CNN and Transformer, achieving global context while maintaining computational efficiency
High-Resolution Processing
Supports 1024x1024 resolution input, suitable for high-precision segmentation tasks
Pre-trained Weights
Provides weights pre-trained on the Cityscapes dataset, ready for direct inference

Model Capabilities

Street scene image segmentation
Pixel-level classification
High-resolution image processing

Use Cases

Autonomous Driving
Road Scene Understanding
Identifies key elements such as roads, pedestrians, and vehicles
Performs well on the Cityscapes dataset
Urban Management
Infrastructure Analysis
Identifies and classifies urban infrastructure such as roads and buildings
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase