Darrell Young mail4y

⚠️ The following instruction is written on 25 Jul 2020 (= old).

4 steps to get a docker running with GPU on Ubuntu 20.04

Make sure time is correct if not reconfigure:
sudo dpkg-reconfigure tzdata

$ gpspipe -w             # log to stdout
$ gpspipe -w -o abc.log  # log to a file

To auto-log on startup, insert the following line to the file /etc/rc.local:

	import rasterio

	# load the geotiff (a DSM in this case) and read the data - only one index in this case
	dsm = rasterio.open(dsm_fname)
	dsm_data = dsm.read()[0]

	# crop the data
	dsm_crop = dsm_data[min_y:min_y+height, min_x:min_x+width]

	# make a copy of the geotiff metadata

	import tensorflow as tf
	import tensorflow.contrib as tfc
	import numpy as np
	import matplotlib.pyplot as plt
	import observations
	from functools import lru_cache

	tf.reset_default_graph()

	train_data, test_data = observations.cifar10('data/cifar',)

	- hosts: all
	become: true
	become_method: sudo
	vars:
	daemon_json:
	bip: 192.168.99.1/24
	default-runtime: nvidia
	runtimes:
	nvidia:
	path: /usr/bin/nvidia-container-runtime

	"""
	Solving FrozenLake8x8 environment using Policy iteration.
	Author : Moustafa Alzantot (malzantot@ucla.edu)
	"""
	import numpy as np
	import gym
	from gym import wrappers


	def run_episode(env, policy, gamma = 1.0, render = False):