Skip to content

Latest commit

 

History

History
42 lines (27 loc) · 1.34 KB

README.md

File metadata and controls

42 lines (27 loc) · 1.34 KB

bird-vision

This is a machine vision project to read information off of a live stream of Final Fantasy Tactics, called FFTBG.

All of this code is written in Python, particulary with opencv and tensorflow. Everything you need should be in the Pipfile except for ffmpeg and stream-link, if you choose to run the code that uses them.

There is a lot of data that goes along with this project, and to download it you will need to make use of Git Large File Storage.

Quick start

Once you have everything installed, and the data downloaded, you can do the following from the project's root directory.

Run tests:

python -m birdvision.scripts.run_tests

Train new models:

python -m birdvision.scripts.train_models --all

Or, if you want to run the web viewer, to visualize test cases:

FLASK_APP=birdvision.web python -m flask run

You can also watch the stream live if you have ffmpeg and stream-link installed:

python -m birdvision.scripts.live_stream

MacOS

You can set up your own RAM Disk like so, useful for mass image downloading / manipulation when you don't necessarily want it to stick around.

diskutil erasevolume HFS+ RAM_Disk_2GB $(hdiutil attach -nomount ram://4000000)