Blog

Showing 6 to 10 of 72 posts.

One can define any policy which is compatible with the Tensorflow-agents library.

In Tensorflow-Agents, we can define our own environments and their associated rewards. 

In the past two days, using the Tensorflow Agents library, I finished writing a code for a DQN agent that finds the optimal Policy for a CartPole problem.

TF-Agents is a library for Reinforcement Learning in TensorFlow. It has very useful functions that help to implement an RL algorithm very fast.

I wrote a code to find the optimal policy for the mountain climb problem with continuous action set.