UperNet-ConvNeXt-Large Open-Source Semantic Segmentation Model - Free for Scenario Parsing Tasks such as ADE20K

Home

Upernet Convnext Large

Developed by smp-hub

UPerNet semantic segmentation model based on ConvNeXt-Large encoder, suitable for scene parsing tasks like ADE20K

Image Segmentation

Safetensors

Open Source License:MIT #Semantic Segmentation #ConvNeXt Architecture #ADE20K Dataset

Downloads 54

Release Time : 4/12/2025

Model Overview

This is a UPerNet architecture semantic segmentation model using ConvNeXt-Large as the encoder, specifically designed for high-precision image semantic segmentation tasks, excelling in parsing complex scenes.

Model Features

High-Performance Encoder

Utilizes ConvNeXt-Large as the encoder, providing powerful feature extraction capabilities

UPerNet Architecture

Employs UPerNet decoder structure to effectively integrate multi-scale features

Pre-trained Support

Provides pre-trained weights for direct inference or fine-tuning

Easy to Use

Integrated with the Albumentations library for convenient preprocessing workflows

Model Capabilities

Image Semantic Segmentation

Scene Parsing

Pixel-Level Classification

Use Cases

Computer Vision

Scene Understanding

Performs pixel-level semantic segmentation of complex scenes to identify different objects and regions

Performs well on the ADE20K dataset

Autonomous Driving

Used for road scene parsing to identify elements like roads, vehicles, and pedestrians

Remote Sensing Image Analysis

Classifies and segments features in satellite or aerial images

🚀 UPerNet Model Card

This is a model card for the UPerNet model, which is used for image segmentation tasks.

🚀 Quick Start

Load trained model

You can quickly start using the pre - trained model through the following steps. Click the button below to open the Colab notebook:

Install requirements.

pip install -U segmentation_models_pytorch albumentations

Run inference.

import torch
import requests
import numpy as np
import albumentations as A
import segmentation_models_pytorch as smp

from PIL import Image

device = "cuda" if torch.cuda.is_available() else "cpu"

# Load pretrained model and preprocessing function
checkpoint = "smp-hub/upernet-convnext-large"
model = smp.from_pretrained(checkpoint).eval().to(device)
preprocessing = A.Compose.from_pretrained(checkpoint)

# Load image
url = "https://huggingface.co/datasets/hf-internal-testing/fixtures_ade20k/resolve/main/ADE_val_00000001.jpg"
image = Image.open(requests.get(url, stream=True).raw)

# Preprocess image
np_image = np.array(image)
normalized_image = preprocessing(image=np_image)["image"]
input_tensor = torch.as_tensor(normalized_image)
input_tensor = input_tensor.permute(2, 0, 1).unsqueeze(0)  # HWC -> BCHW
input_tensor = input_tensor.to(device)

# Perform inference
with torch.no_grad():
    output_mask = model(input_tensor)

# Postprocess mask
mask = mask.argmax(1).cpu().numpy()  # argmax over predicted classes (channels dim)

✨ Features

Model init parameters

The following are the initialization parameters of the model:

model_init_params = {
    "encoder_name": "tu-convnext_large",
    "encoder_depth": 5,
    "encoder_weights": None,
    "decoder_channels": 512,
    "decoder_use_norm": "batchnorm",
    "in_channels": 3,
    "classes": 150,
    "activation": None,
    "upsampling": 4,
    "aux_params": None
}

Dataset

The model uses the following dataset: Dataset name: ADE20K

📚 Documentation

More Information

Library: https://github.com/qubvel/segmentation_models.pytorch
Docs: https://smp.readthedocs.io/en/latest/

This model has been pushed to the Hub using the PytorchModelHubMixin

📄 License

This project is licensed under the MIT 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