SegFormer Trainer Test Bis Open-Source Image Segmentation Model - Free Deployment for Sidewalk Semantic Segmentation

Home

Segformer Trainer Test Bis

Developed by nielsr

An image segmentation model fine-tuned from nvidia/mit-b0 for sidewalk semantic segmentation tasks

Image Segmentation

Transformers

Open Source License:Apache-2.0 #Sidewalk Semantic Segmentation #Traffic Scene Parsing #MIT-B0 Fine-tuning

Downloads 14

Release Time : 4/19/2022

Model Overview

This model is a fine-tuned version of nvidia/mit-b0 on the segments/sidewalk-semantic dataset, primarily used for semantic segmentation tasks in urban road scenes.

Model Features

Lightweight Architecture

Lightweight Transformer architecture based on mit-b0, suitable for resource-constrained environments

Multi-class Segmentation

Supports recognition of 40+ road scene elements including roads, sidewalks, buildings, etc.

Fine-tuning Optimization

Specially optimized on the segments/sidewalk-semantic dataset, suitable for urban road scenes

Model Capabilities

Semantic Image Segmentation

Road Scene Analysis

Multi-class Recognition

Use Cases

Smart City

Sidewalk Infrastructure Analysis

Automatically identify and classify various infrastructure elements in urban sidewalk areas

Can recognize elements such as roads, sidewalks, buildings, etc.

Road Safety Monitoring

Detect traffic signs, pedestrians, and other elements on roads to assist safety monitoring

Can detect targets such as traffic signs and pedestrians

Autonomous Driving

Road Scene Understanding

Provide semantic understanding of road scenes for autonomous driving systems

Can identify key elements such as roads, sidewalks, vehicles, etc.

🚀 segformer-trainer-test-bis

This model is a fine - tuned version of [nvidia/mit - b0](https://huggingface.co/nvidia/mit - b0) for image segmentation, offering specific performance metrics on the evaluation set.

🚀 Quick Start

This model, segformer - trainer - test - bis, is a fine - tuned version of [nvidia/mit - b0](https://huggingface.co/nvidia/mit - b0) on the segments/sidewalk - semantic dataset. It achieves the following results on the evaluation set:

Metric	Value
Loss	1.3784
Mean Iou	0.1424
Mean Accuracy	0.1896
Overall Accuracy	0.7288
Accuracy Unlabeled	nan
Accuracy Flat - road	0.6651
Accuracy Flat - sidewalk	0.9129
Accuracy Flat - crosswalk	0.0
Accuracy Flat - cyclinglane	0.5829
Accuracy Flat - parkingdriveway	0.0184
Accuracy Flat - railtrack	0.0
Accuracy Flat - curb	0.0
Accuracy Human - person	0.0
Accuracy Human - rider	0.0
Accuracy Vehicle - car	0.8322
Accuracy Vehicle - truck	0.0
Accuracy Vehicle - bus	0.0
Accuracy Vehicle - tramtrain	0.0
Accuracy Vehicle - motorcycle	0.0
Accuracy Vehicle - bicycle	0.0
Accuracy Vehicle - caravan	0.0
Accuracy Vehicle - cartrailer	0.0
Accuracy Construction - building	0.8930
Accuracy Construction - door	0.0
Accuracy Construction - wall	0.0025
Accuracy Construction - fenceguardrail	0.0
Accuracy Construction - bridge	0.0
Accuracy Construction - tunnel	0.0
Accuracy Construction - stairs	0.0
Accuracy Object - pole	0.0008
Accuracy Object - trafficsign	0.0
Accuracy Object - trafficlight	0.0
Accuracy Nature - vegetation	0.8552
Accuracy Nature - terrain	0.8507
Accuracy Sky	0.8336
Accuracy Void - ground	0.0
Accuracy Void - dynamic	0.0
Accuracy Void - static	0.0
Accuracy Void - unclear	0.0
Iou Unlabeled	nan
Iou Flat - road	0.4712
Iou Flat - sidewalk	0.7651
Iou Flat - crosswalk	0.0
Iou Flat - cyclinglane	0.5216
Iou Flat - parkingdriveway	0.0178
Iou Flat - railtrack	0.0
Iou Flat - curb	0.0
Iou Human - person	0.0
Iou Human - rider	0.0
Iou Vehicle - car	0.5696
Iou Vehicle - truck	0.0
Iou Vehicle - bus	0.0
Iou Vehicle - tramtrain	0.0
Iou Vehicle - motorcycle	0.0
Iou Vehicle - bicycle	0.0
Iou Vehicle - caravan	0.0
Iou Vehicle - cartrailer	0.0
Iou Construction - building	0.4716
Iou Construction - door	0.0
Iou Construction - wall	0.0024
Iou Construction - fenceguardrail	0.0
Iou Construction - bridge	0.0
Iou Construction - tunnel	0.0
Iou Construction - stairs	0.0
Iou Object - pole	0.0008
Iou Object - trafficsign	0.0
Iou Object - trafficlight	0.0
Iou Nature - vegetation	0.6813
Iou Nature - terrain	0.5513
Iou Sky	0.7873
Iou Void - ground	0.0
Iou Void - dynamic	0.0
Iou Void - static	0.0
Iou Void - unclear	0.0

📚 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: 5e - 05
train_batch_size: 16
eval_batch_size: 16
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
num_epochs: 10.0

Training results

More information needed

Framework versions

Transformers 4.19.0.dev0
Pytorch 1.11.0+cu113
Datasets 2.0.0
Tokenizers 0.11.6

📄 License

This model 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