search

Unity-Technologies / ml-agents

The Unity Machine Learning Agents Toolkit (ML-Agents) is an open-source project that enables games and simulations to serve as environments for training intelligent agents using deep reinforcement learning and imitation learning.
https://unity.com/products/machine-learning-agents
Other
17.04k stars 4.15k forks source link

Integration of TFLite in Unity TF-Sharp for optimised mobile performance #1359

Closed chetu3319 closed 5 years ago

chetu3319 commented 5 years ago

Hi Community,

I'm trying to integrate an optimized graph of Deeplab: https://ai.googleblog.com/2018/03/mobile-real-time-video-segmentation.html into unity for a real-time video segmentation.

I got some success with simply TF Deeplab model in unity, but on running it, the performance is like 4-10 seconds per frame, which is quite not what the application requirements.

I'm looking intoTFLite unity integration as mentioned on the Github repo: https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/lite/experimental/examples/unity/TensorFlowLitePlugin

Target device: Galaxy Note 9 OS: Android Building on : Windows10 OS Unity Version : Unity 2018.2.10f1

It's giving a real hard time, to follow along and to build necessary plugins. Help on this will be appreciated. -Chaitanya

chetu3319 commented 5 years ago

@awjuliani, Any update which you can share?

awjuliani commented 5 years ago

Hi @chetu3319

We will be switching to an internal Unity inference solution for ML-Agents v0.7. As such we won't be supporting TF# or TFLite going forward.

github-actions[bot] commented 3 years ago

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.