Skip to content

Instantly share code, notes, and snippets.


Jeremiah Johnson jwjohnson314

View GitHub Profile
jwjohnson314 /
Last active Jun 6, 2016 — forked from karpathy/
Training a Neural Network ATARI Pong agent with Policy Gradients from raw pixels
#!/usr/bin/env python
Trains an agent with (stochastic) Policy Gradients on Pong. Uses OpenAI Gym.
Python3 port
import numpy as np
import pickle
import gym
# hyperparameters
You can’t perform that action at this time.