Custom instance segmentation pytorch

Language and Libraries

Problem statement

Solution Proposed

In this project, you have learned how to create your own training pipeline for instance segmentation models, on a custom dataset. For that, we wrote a torch.utils.data.Dataset class that returns the images and the ground truth boxes and segmentation masks. We also leveraged a Mask R-CNN model pre-trained on COCO train2017 in order to perform transfer learning on this new dataset.

Dataset Used

Penn-Fudan Database for Pedestrian Detection and Segmentation. It contains 170 images with 345 instances of pedestrians, and we will use it to illustrate how to use the new features in torchvision in order to train an instance segmentation model on a custom dataset.

Model Used

In this project, we will be using Mask R-CNN, which is based on top of Faster R-CNN. Faster R-CNN is a model that predicts both bounding boxes and class scores for potential objects in the image.

How to run?

Step 1: Clone the repository

git clone my repository 

Step 2- Create a conda environment after opening the repository

conda create -p env python=3.8 -y

conda activate env

Step 3 - Install the requirements

pip install -r requirements.txt

Step 4 - Install Google Cloud Sdk and configure

For Windows

https://dl.google.com/dl/cloudsdk/channels/rapid/GoogleCloudSDKInstaller.exe

For Ubuntu

sudo apt-get install apt-transport-https ca-certificates gnupg

echo "deb [signed-by=/usr/share/keyrings/cloud.google.gpg] https://packages.cloud.google.com/apt cloud-sdk main" | sudo tee -a /etc/apt/sources.list.d/google-cloud-sdk.list

curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key --keyring /usr/share/keyrings/cloud.google.gpg add -

sudo apt-get update && sudo apt-get install google-cloud-cli

gcloud init

Before running server application make sure your Google Cloud Storage bucket is available

Step 5 - Run the application server

python app.py

Run locally

1. Check if the Dockerfile is available in the project directory

2. Build the Docker image

docker build -t track . 

3. Run the Docker image

docker run -d -p 8080:8080 <IMAGEID>

4. Open docker image in interactive model

docker exec -ti <IMAGEID> bash

5. Authenticate GCloud

gcloud auth login

6. Authenticate default application

gcloud auth application-default login

👨‍💻 Tech Stack Used

1. Python
2. Pytorch
3. Docker

🌐 Infrastructure Required.

1. Google Cloud Storage
2. Google Compute Engine
3. Google Artifact Registry
4. Circle CI

src is the main package folder which contains

Artifact : Stores all artifacts created from running the application

Components : Contains all components of this project

• DataIngestion
• DataTransformation
• ModelTrainer
• ModelEvaluation
• ModelPusher

Custom Logger and Exceptions are used in the project for better debugging purposes.

Conclusion

=====================================================================