Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/codeamt/fastefficientcovidnet

CXR classification API with pre-trained EfficientNet-b1 and Fast.ai
https://github.com/codeamt/fastefficientcovidnet

Last synced: about 2 months ago
JSON representation

CXR classification API with pre-trained EfficientNet-b1 and Fast.ai

Awesome Lists containing this project

README

        

# Fast Efficient CovidNet - An End-to-End Pipeline
This GitHub Repository contains my final project for Udacity's Machine Learning Engineer Nanodegree

## About
This is a Chest X-Ray (CXR) classification API. Building on previous work of [[1]](https://arxiv.org/pdf/2003.09871v3.pdf), the CovNet model for this ML project utilizes a pre-trained EfficientNet-b1 to extract features and a fine-tuned Fast.ai classifier to differentiate between infection classes (Normal, Viral Pneumonia, or COVID-19) with 95% test accuracy.

## Repo Contents:
- [mle-capstone-data (submodule)](https://github.com/codeamt/mle-capstone-data/tree/41128aef7f98d479fafc6c946f804b5e52d8c89f)
- [mle-capstone-modeling (submodule)](https://github.com/codeamt/mle-capstone-modeling/tree/6d2469b585b6d0d7535a33e734453b07846a5cb6)
- [mle-capstone-deployment (submodule)](https://github.com/codeamt/mle-capstone-deployment/tree/16741ab5f2c4670dff3c962cc0ba5b9580e8b9b2)
- [project propsal (pdf)](https://github.com/codeamt/udacity-mle-capstone-project/blob/master/proposal.pdf)
- [project report (pdf)](https://github.com/codeamt/udacity-mle-capstone-project/blob/master/report.pdf)
- [README.md (here)]()

## Instructions:
Each submodule provides an .ipynb file for a detailed walkthrough of that project phase.

In the data submodule, to generate the COVIDx dataset, you'll neet a kaggle account. If you don't have an account, you can create one at [kaggle.com](https://www.kaggle.com/). From there, from the top right menu, visit My Account > API > create New API Token > ... to get a JSON hardfile of your API Token.

In the deployment submodule, if you'd like to run the quick start demo notebook, you'll also need to create a [ngrok]() account and API token.