Spaces:
Sleeping
Sleeping
| <html> | |
| <head> | |
| <title>DeepCubeA</title> | |
| <meta name="viewport" content="width=device-width, initial-scale=1.0"> | |
| <!-- Bootstrap --> | |
| <link href="//cdnjs.cloudflare.com/ajax/libs/twitter-bootstrap/3.3.7/css/bootstrap.min.css" rel="stylesheet"> | |
| <link rel="stylesheet" href="/static/main.css"> | |
| <!--[if lt IE 9]> | |
| <script src="//cdnjs.cloudflare.com/ajax/libs/html5shiv/3.7.3/html5shiv.min.js"></script> | |
| <script src="//cdnjs.cloudflare.com/ajax/libs/respond.js/1.4.2/respond.min.js"></script> | |
| <![endif]--> | |
| <!-- 移除重复的引入代码 --> | |
| <link href="//cdnjs.cloudflare.com/ajax/libs/twitter-bootstrap/3.3.7/css/bootstrap.min.css" rel="stylesheet"> | |
| <link rel="stylesheet" href="/static/main.css"> | |
| <link rel="shortcut icon" href="static/favicon.ico"> | |
| <meta name="twitter:card" content="summary_large_image"> | |
| <meta name="twitter:title" content="DeepCubeA: Solve the Rubik's cube With Deep Learning"> | |
| <meta name="twitter:description" content="DeepCubeA trains itself using deep learning and reinforcement learning. It then uses A* search to solve the cube."> | |
| <meta name="twitter:image" content="https://www.ics.uci.edu/~fagostin/cube.jpg"> | |
| </head> | |
| <body> | |
| <div class="container"> | |
| <center> | |
| <h1>Solve the Rubik's Cube Using Deep Learning</h1> | |
| <div id="program_container" class="container"> | |
| <div id="solution_container" class="container"> | |
| <button disabled id="first_state" type="button"><<</button> | |
| <button disabled id="prev_state" type="button"><</button> | |
| <button disabled id="next_state" type="button">></button> | |
| <button disabled id="last_state" type="button">>></button> | |
| <h4 id="solution_text">Solution:</h4> | |
| </div> | |
| <div id="cube_div" class="container" style="padding: 60px"> | |
| <section class="cube_container"> | |
| <div id="cube"> | |
| <figure id="front" class="front"> | |
| </figure> | |
| <figure id="back" class="back"> | |
| </figure> | |
| <figure id="right" class="right"> | |
| </figure> | |
| <figure id="left" class="left"> | |
| </figure> | |
| <figure id="top" class="top"> | |
| </figure> | |
| <figure id="bottom" class="bottom"> | |
| </figure> | |
| </div> | |
| </section> | |
| </div> | |
| <br> | |
| <div id="button_container" class="container"> | |
| <button disabled id="scramble" type="button">Scramble</button> | |
| <button disabled id="solve" type="button">Solve!</button> | |
| <button disabled id="clear" type="button">Clear</button> | |
| </div> | |
| </div> | |
| </center> | |
| </div> | |
| <br> | |
| <center> | |
| <h4> | |
| Use the mouse to turn the cube. | |
| </h4> | |
| <h4> | |
| Turn the faces with the U/D/L/R/B/F keys. | |
| Hold shift to turn faces couter-clockwise. | |
| </h4> | |
| <h4> | |
| Press scramble to randomly scramble the cube. | |
| Press solve to solve the cube using deep learning! | |
| </h4> | |
| <h5> | |
| <i>Note: Due to model capacity limitations, there might be cases where the model cannot solve in reasonable time.</i> | |
| </h5> | |
| </center> | |
| <br> | |
| <div class="well well-lg"> | |
| <center> | |
| <a href="https://github.com/forestagostinelli/DeepCubeA">Code (Latest)</a> - <a href="https://github.com/xiaofeng218/DeepcubeA">Code (Reproduce)</a> | |
| <br> | |
| <br> | |
| <span style='color:black'>If you find this research useful to your work, please cite the following papers:</span><br><br> | |
| <span class="paper-title"><a href="static/files/SolvingTheRubiksCubeWithDeepReinforcementLearningAndSearch_Final.pdf">Solving the Rubik's Cube with Deep Reinforcement Learning and Search</a></span><br> | |
| <!--span class="paper-title"><a href="https://www.nature.com/articles/s42256-019-0070-z.epdf?shared_access_token=-pCSsZa_J9bM8VyXLZLRctRgN0jAjWel9jnR3ZoTv0Osb8UCgUm5AQaSCMHWqWzsyV3KBcb13SAW-9IL1pAGd1HcSk40JSEjhoaBAi0ePvYh_5Dul6LvK0oJY1KI0ULo9O9HCut_y7aCTc93Th8m5g%3D%3D">Solving the Rubik's Cube with Deep Reinforcement Learning and Search</a></span><br--> | |
| <span class="paper-authors">Forest Agostinelli*, Stephen McAleer*, Alexander Shmakov*, Pierre Baldi</span><br> | |
| <span class="paper-journal"><i>Nature Machine Intelligence</i>, Volume 1</span>, | |
| <span class="paper-year">2019</span> | |
| <br> | |
| <br> | |
| <span class="paper-title"><a href="https://openreview.net/pdf?id=Hyfn2jCcKm">Solving the Rubik's Cube with Approximate Policy Iteration</a></span><br> | |
| <span class="paper-authors">Stephen McAleer*, Forest Agostinelli*, Alexander Shmakov*, Pierre Baldi</span><br> | |
| <span class="paper-journal"><i>In Proceedings of the 7th International Conference on Learning Representations (ICLR)</i></span>, | |
| <span class="paper-year">2019</span> | |
| </center> | |
| </div> | |
| <center> | |
| <h5> | |
| Note: For ease of maintenance, this updated version solves the cube using CPUs instead of a GPU. | |
| Therefore, solve times will be around 20 seconds instead of 1 second. | |
| </h5> | |
| </center> | |
| <script src="https://ajax.googleapis.com/ajax/libs/jquery/3.2.1/jquery.min.js"></script> | |
| <script src="/static/main.js"></script> | |
| <script src="//cdnjs.cloudflare.com/ajax/libs/twitter-bootstrap/3.3.7/js/bootstrap.min.js"></script> | |
| </body> | |
| </html> | |