muupan

"""A training script of Soft Actor-Critic on OpenAI Gym Mujoco environments.

This script follows the settings of https://arxiv.org/abs/1812.05905 as much
as possible.
"""
import argparse
from distutils.version import LooseVersion
import functools
import logging
import sys

muupan

How to build DeepMind Lab on macOS Mojave (as of 2019/02/01)

1. Clone macos branch of https://github.com/deepmind/lab

2. Read https://github.com/deepmind/lab/blob/macos/README.md and install dependencies.

3. Edit WORKSPACE to correct paths of glib_archive, sdl_system, and python_system.

• For glib_archive and sdl_system, you need to specify their correct paths.

muupan

#!/bin/sh

set -e

# Download all the pdfs
wget -nc http://www.ee.iitm.ac.in/~krishnaj/EE5110_files/notes/lecture1_set_theory.pdf
wget -nc http://www.ee.iitm.ac.in/~krishnaj/EE5110_files/notes/lecture2_Realanalysis.pdf
wget -nc http://www.ee.iitm.ac.in/~krishnaj/EE5110_files/notes/lecture3_cardinality.pdf
wget -nc http://www.ee.iitm.ac.in/~krishnaj/EE5110_files/notes/lecture4_probability_spaces.pdf
wget -nc http://www.ee.iitm.ac.in/~krishnaj/EE5110_files/notes/lecture5_properties%20of%20prob%20measures.pdf

muupan

from timeit import default_timer as timer

import chainer
from chainer import cuda
from chainer import function
import chainer.functions as F
from chainer import utils
from chainer.utils import type_check
import cupy

muupan

Guided Cost Learning: Deep Inverse Optimal Control via Policy Optimization
Doubly Robust Off-policy Value Evaluation for Reinforcement Learning
Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning
Learning Simple Algorithms from Examples
Stability of Controllers for Gaussian Process Forward Models
Smooth Imitation Learning for Online Sequence Prediction
On the Analysis of Complex Backup Strategies in Monte Carlo Tree Search
Benchmarking Deep Reinforcement Learning for Continuous Control
Cumulative Prospect Theory Meets Reinforcement Learning: Prediction and Control
Why Most Decisions Are Easy in Tetris—And Perhaps in Other Sequential Decision Problems, As Well

muupan

# Move the current window left/right by Ctrl+Shift+Left/Right
bind-key -n C-S-Left swap-window -t -1
bind-key -n C-S-Right swap-window -t +1

# Move to the prev/next window by Shift+Left/Right
bind-key -n S-Left prev
bind-key -n S-Right next

muupan

# CMakeLists.txt that makes all the header and source files visible from Qt Creator

cmake_minimum_required(VERSION 2.8)
execute_process(
  COMMAND find . -name *.cpp -or -name *.hpp -or -name *.h -or -name *.cc -or -name *.hh
  COMMAND tr "\n" ";"
  OUTPUT_VARIABLE QTCREATOR_SRCS
  OUTPUT_STRIP_TRAILING_WHITESPACE
)
add_custom_target(qtcreator SOURCES ${QTCREATOR_SRCS})

muupan

#include <array>
#include <iostream>
#include <random>

int main() {
    constexpr std::array<double, 3> fitness_values = {1, 1.5, 2};
    // Constructor is O(n)
    std::discrete_distribution<std::size_t> dist(fitness_values.begin(), fitness_values.end());
    std::array<size_t, 3> table;
    table.fill(0);

muupan

layers {
  name: "frames_input_layer"
  type: MEMORY_DATA
  top: "frames"
  top: "dummy1"
  memory_data_param {
    batch_size: 32
    channels: 4
    height: 84
    width: 84

muupan

#include <iostream>
#include <memory>
#include <random>
#include <caffe/caffe.hpp>
#include <glog/logging.h>

int main(int argc, char** argv) {
  // glogの初期化
  google::InitGoogleLogging(argv[0]);