AI Startup with PyTorch
Overview
This is the code for this video on Youtube by Siraj Raval on building an Finance Startup. You should use this code and the video as a boilerplate for building your own AI startup.
#AISTARTUPCHALLENGE
Due date is May 15 2019 Submit your entry here.
Call for Pull Requests
This code compiles and runs, but its far from complete. I'm accepting pull requests, i encourage you to do so, it will be a service to everyone who watches the video.
- Integrate Firebase with the login in a more intuitive way
- Finish integrating AnyCharts so it pulls from the data directly
- Add more Tensorflow models for serving
- Add more instructions for the dialogflow agent
- Make better investments using Alpaca, a better trading strategy than just using sentiment analysis
- Speaking of Sentiment Analysis, make that work in a more efficient way
Dependencies
- Firebase
- Plaid
- Tensorflow Lite
- DialogFlow
- AnyChart
Instructions
Download Android Studio then import this project. Since I've placed the dependencies in the build.gradle file, it should auto-download them for you, which is awesome. It needs more features, and better integrated dataflow. I could've put more time into it, but I'm on to the next one. The model was too big to upload to GitHub. Place any .pb model into the app/src/main/assets folder and Tensorflow lite will detect it so it can be used for inference.
Credits
So many people. I stand on the shoulders of giants. Respect to all developers.