This is a wiki post - feel free to edit to add links from the lesson or other useful info.
Lesson resources
- Lesson video (private)
- Covid-19 masks video (public)
- fastbook chapter 2
- fastbook Chapter 2 questionnaire solutions - feel free to contribute!
- Non-beginner discussion
- Part 1 (2020) - Weekly Beginner Only Review and Q&A
- Anki deck for Lessons 1 and 2
Links from lesson
- Masks research
- masks4all - from Czech Republic
- Publication “The ASA Statement on p -Values” & article " Null Hypothesis Significance Testing Never Worked"
- Designing great data products
- Bing Image Search API
Deployment guides
- Deploying the Bear Detector app with Flask + Gunicorn + Domain name with SSL certificate, plus Dockerizing the app: https://github.com/fastai/course-v3/blob/master/docs/deployment_flask_gunicorn_docker.md (just sent pull request to correct missing /docs in link to image oops)
- Deployment Platform: SeeMe.ai - the AI marketplace: Easily deploy and share your Fastai (v1 and v2) models on the web, mobile, Python SDK, … We have added support for
fastai2
and will continue to support it throughout its development to a stable version.
We created afastai2
deployment quick guide, taking you through all the steps of training, deploying, using, and sharing your model with your friends and the world.
Additional links: Deployment platform: Seeme.ai - wiki topic | Fastai v1 quick guide - Deployment Platform: Amazon SageMaker. Follow the steps outlined in the following post to train and deploy your fastai2 model with Amazon SageMaker.
- Deployment Platform: Azure Functions. Follow the steps outlined in this Github repo to deploy a trained fastai2 model on Azure cloud’s serverless infrastructure service – the Azure Functions. Welcome your feedback and questions on this topic thread.
Deployment examples
- Bear Detector web app: a Starlette example for deployment in fastai2
- Derotate: a code template for image-to-class and image-to-image fastai2 deployments with web app, mobile app (in Flutter), and a notebook showing how to call the model remotely as an API.
- Bear detector web app deployed with flask: example of the bear detector app deployed with fast.ai v2, flask, gunicorn and nginx