Skip to content

Instantly share code, notes, and snippets.

View lemonteaa's full-sized avatar

Lemontea lemonteaa

View GitHub Profile
@lemonteaa
lemonteaa / Keras-RL+PPO+Test.py
Last active November 26, 2017 14:37
Jupyter Notebook for developing my PPO integration into keras-rl
# coding: utf-8
# # Getting Started
#
# First we need to figure out the details of keras and gyms input spec, especially the tensor's shape
#
# Note that for a numpy (**n-d**imensional) array, it has a `shape` attr which is a python tuple. For example, `s = (2,6,3,)` represents a $2 \times 6 \times 3$ tensor.
#
# Also note that `np.array` can convert a python array into a numpy array, with the *obvious* interpretation. For example,