Are there any good tutorials about training RL agent from raw pixels using PyTorch?


Is there any good tutorials about training reinforcement learning agent from raw pixels using PyTorch?

I don't understand the official PyTorch tutorial. I want to train the agent on the atari breakout environment. Unfortunately, I failed to train the agent on the RAM version. Now, I am looking for a way to train the agent from raw pixels.

dato nefaridze

Posted 2020-06-03T11:23:44.193

Reputation: 464



I am using Jetson reinforcement for my quadcopter reinforcement in simulation. Maybe it will help you. because you can create AI agents that can learn from the interactive environment, gather experience, and system of reward with deep RL. You can also use end-to-end neural networks that translate raw pixels into action as per your need and use that RL-trained agent to complete complex tasks.

The best thing is you can easily transfer RL-agent which simulated in the simulator to real-world robots. and We are using multiple Nano's easily to perform the complex tasks of navigation and co-ordination of Our Quad-copter.

Webinar of Deep RL on Jetson

Google group to get help for Deep RL on jetson nano

Hope it helps you.

Hiren Namera

Posted 2020-06-03T11:23:44.193

Reputation: 125

The question was about tutorials, so you should at least explain why your suggestion could help to solve the issue. – nbro – 2020-06-03T22:43:10.097


I always found that towards data science articles can be a good source for code for these types of problems. They usually have a git repo with everything you need and walk through the less obvious steps in the article. This article may be of interest to you.

David Ireland

Posted 2020-06-03T11:23:44.193

Reputation: 1 942

ok thanks I am going to check that – dato nefaridze – 2020-06-03T14:11:58.360