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
16.93k stars 4.14k forks source link

The hyper-parameter Gamma could not be found for the <class 'mlagents.trainers.ppo.trainer.PPOTrainer'> trainer of brain RubixLearner. #2463

Closed skamal16 closed 5 years ago

skamal16 commented 5 years ago

I get the error described in the heading when I try to run ml-agents-learn and then press play. The console window gets stuck and the unity editor window also seems to get stuck. I have to Ctrl-C to interrupt the console to do anything else.

Note: This is the first time I am running ML-Agents so I am probably missing something obvious. Any guidance would be appreciated.

Update: It's not just my scene that this doesn't work on, it is the same whether I run on any of the example scenes. So it shouldn't be related to the way I've set my learning brain.

Update: Reinstalling ml-agents and reimporting ml agent assets seems to have solved the problem. I'm assuming the problem had something to due with a mismatch of versions of ml agents of where the trainer file was and of the brain assets I was using in my project. Anyways I'll be closing this now.

Coookiez commented 5 years ago

I'm using the newest version cloned from the repo so when I deleted the ML-Agents folder and reimported it - nothing changed. Any other suggestions on how to fix that problem?

skamal16 commented 5 years ago

Your assets may be of the newest version, but when last did you install/update your ml agents environment? I think my main issue was the need to update my ml agents environment, in which case I just deleted it and installed it from scratch. If this doesn't fix it then I'm not sure what the problem may be.

Coookiez commented 5 years ago

I did it like 2-3 days prior to running the agents.

UPDATE: I did as you said and that error's gone, however I get this now: "Key Error: 'reward_signals'" Any idea what could that mean?

image

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.