This repository hosts a customized PPO based agent for Carla. The goal of this project is to make it easier to interact with and experiment in Carla with reinforcement learning based agents -- this, by wrapping Carla in a gym like environment that can handle custom reward functions, custom debug output, etc.
I would like to ask about the VAE training, I only used your RGB data set and got the following data roughly, but will the loss be too big.
Also, I would like to ask about this plot obtained using vae_plots.py. Can I think of it as 64 is 64 features, and then the ordinate on the same scale is a visualization of different features on the same image? And what is this [-10,10] scale down here?
I would like to ask about the VAE training, I only used your RGB data set and got the following data roughly, but will the loss be too big.
Also, I would like to ask about this plot obtained using vae_plots.py. Can I think of it as 64 is 64 features, and then the ordinate on the same scale is a visualization of different features on the same image? And what is this [-10,10] scale down here?
Thanks.@bitsauce