Crystal Vision is an AI able to predict the result of a LOL game, from and only from the draft time.
The model is currently ~70% accurate.
I am an IA beginner. This is my first project about AI, neural network, using Python, with TensorFlow.
All the data used to train the model were scraped using the official RIOT API (~80 000 games...)
If you reach to have better result or if you have improvements to give. You can pull your changes because this interests me a lot.
- Test league of legends local API
- Optimization + better code
- Data analytics to train better the IA's model
- Update dependencies (TensorFlow...)
You can easly contact me on Discord, if you got question or improvement...
@forwarzz