S

Segformer B0 Finetuned Cityscapes 512 1024

Developed by nvidia
SegFormer is a Transformer-based semantic segmentation model fine-tuned on the Cityscapes dataset, suitable for road scene segmentation tasks.
Downloads 1,111
Release Time : 3/2/2022

Model Overview

This model adopts a hierarchical Transformer encoder and lightweight all-MLP decoder head architecture, specifically optimized for 512x1024 resolution Cityscapes dataset for semantic segmentation tasks.

Model Features

Hierarchical Transformer architecture
Adopts a hierarchical Transformer encoder to effectively capture multi-scale features
Lightweight MLP decoder head
Uses a lightweight all-MLP decoder head to maintain efficient inference speed
Cityscapes optimization
Specifically fine-tuned and optimized for the Cityscapes road scene dataset

Model Capabilities

Road scene semantic segmentation
High-resolution image processing (512x1024)
Multi-class pixel-level classification

Use Cases

Intelligent transportation
Autonomous driving scene understanding
Identifies traffic elements such as roads, pedestrians, and vehicles
Example images demonstrate accurate segmentation of road scenes
Urban digitization
Street view image analysis
Performs semantic segmentation of urban street scenes to support city planning
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase