Serve MMSegmentation

Deploy model as REST API service

apps neural network images videos semantic segmentation segmentation & tracking serve

Run in Supervisely View at GitHub

Readme Releases 78

Serve MMSegmentation

Overview • How To Run • How To Use • How To Use Your Trained Model Outside Supervisely • Related Apps • Acknowledgment

GitHub release (latest SemVer)

Overview

Serve MMSegmentation model as Supervisely Application. MMSegmentation is an open source semantic segmentation toolbox based on PyTorch. Learn more about MMSegmentation and available models here.

Model serving allows to apply model to image (URL, local file, Supervisely image id) with 2 modes (full image, image ROI). Also app sources can be used as example how to use downloaded model weights outside Supervisely.

Application key points:

Serve custom and MMSegmentation models
Deployed on GPU

Available models

Supported backbones:

Supported methods:

How to Run

Step 1. Add Serve MMSegmentation app to your team from Ecosystem

Serve MMSegmentation
app

Deploy model as REST API service

Step 2. Run the application from Plugins & Apps page

How to Use

Pretrained models

Step 1. Select architecture, pretrained model and press the Serve button

Step 2. Wait for the model to deploy

Custom models

Model and directory structure must be acquired via Train MMSegmentation app or manually created with the same directory structure

How To Use Your Trained Model Outside Supervisely

You can use your trained models outside Supervisely platform without any dependencies on Supervisely SDK. You just need to download config file and model weights (.pth) from Team Files, and then you can build and use the model as a normal model in mmsegmentation. See this Jupyter Notebook for details.

Related apps

You can use served model in next Supervisely Applications ⬇️

Train MMSegmentation - app allows to play with different inference options, monitor metrics charts in real time, and save training artifacts to Team Files.

Train MMSegmentation GPU
app

Dashboard to configure, start and monitor training

Apply NN to images project - app allows to play with different inference options and visualize predictions in real time. Once you choose inference settings you can apply model to all images in your project to visually analyse predictions and perform automatic data pre-labeling.

Apply NN to Images Project
app

NN Inference on images in project or dataset

Apply NN to Videos Project - app allows to label your videos using served Supervisely models.

Apply NN to Videos Project
app

Predictions on every frame are combined with BoT-SORT/DeepSort into tracks automatically

NN Image Labeling - integrate any deployd NN to Supervisely Image Labeling UI. Configure inference settings and model output classes. Press Apply button (or use hotkey) and detections with their confidences will immediately appear on the image.