Unity In Editor Deep Learning Tools. Using KerasSharp, TensorflowSharp, Unity MLAgent. In-Editor training and no python needed.
It is an extension of Unity ML agent for deep learning, primarily reinforcement learning, with in-editor/in-game training support. It also provides interface for another optimization algorithms such as MAES.
It uses a modified version of KerasSharp and TensorflowSharp as the backend, which usesTensorflow c++ lib. No python is needed for model building/evaluation/training. You can even build a standalone(an actual playable game!) with training capability.
This repo is made for Aalto University's Intellicent Computational Media course. The course includes two parts: Audio(Python) and Games(Unity and Python), and this repo contains the main materials for the Unity part of the course. This repo is a remake based on the original materials, which are made with CNTK. It will also be part of my master's thesis.
We might or might not update this repository. But it will try to keep it up with Unity ML-Agents at least.
Possible future plans/contributions:
MIT.