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
17.17k stars 4.16k forks source link

Support adding observations with Matrix in the future? #845

Closed DavidLining closed 5 years ago

DavidLining commented 6 years ago

I plan to use 3D convolution nerual network with PPO for my project recently, I don't know if it can works. haha~ However, it seems that ML-Agents only support adding one-dimensional array observation to the vector of observations of the agent. Will the ML-Agents team support adding multi-dimensional array obvervation in the future?

Thanks.

MarcoMeter commented 6 years ago

As of now, ML-Agents supports camera image observations, which are fed to a convolutional layer.

DavidLining commented 6 years ago

@MarcoMeter yes, but it's an 2D convolutional layer. It's better if ML-Agents can support 3D inputs in the future.

awjuliani commented 6 years ago

Hi All,

We plan to add more variable observation space types in the future, including matrices.

mmattar commented 5 years ago

Thank you for your comments. We have this in our roadmap to complete in the future. We will communicate when this is available in a later release.

Amr311 commented 4 years ago

Hello everyone,

Is this available now? to pass a 2D array as an observation?

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.