search

araffin / learning-to-drive-in-5-minutes

Implementation of reinforcement learning approach to make a car learn to drive smoothly in minutes
https://towardsdatascience.com/learning-to-drive-smoothly-in-minutes-450a7cdb35f4
MIT License
287 stars 87 forks source link

Input to the Donkey Simulator #2

Closed kk2491 closed 5 years ago

kk2491 commented 5 years ago

Hi,

Thanks for the repository.

I have downloaded the code and started running the simulator. Could you please help me in understanding the below queries.

  1. What are the inputs to the algorithm?
  2. If camera is the input, how can I extract the camera images?

Thank you, KK

araffin commented 5 years ago

What are the inputs to the algorithm?

Input to the VAE is an image, input to the RL algo is a vector (encoded image using the VAE).

If camera is the input, how can I extract the camera images?

I don't really understand that question... Simulator directly sends the image.

kk2491 commented 5 years ago

@araffin Thanks for the reply. I wanted to capture the images during the training process.

araffin commented 5 years ago

There is already the recorder wrapper for that.

kk2491 commented 5 years ago

@araffin Yep.. Got you.. Thank you so much for the swift response. I have been working on this project will post here if any queries.