(44803)پروژه big data /

رشته big data درسم میشه همون big data تقریبا ده روزی وقت هست این پروژه دانشجوییو لطف میکنید تا انتها مطالعه کنید، ببینید که تیمتون میتونن برام انجام بدن یا خیر. از مراحل کار اسکرین شات میخام، همچنین مطابق چیزی که کفته شده ریپورت کامل و خود پروژه Summary: 1. Read the article on Reinforcement Learning from Andrej Karpathy: http://karpathy.github.io/2016/05/31/rl (http://karpathy.github.io/2016/05/31/rl/)/. 2. Download the Python code (130 lines) from https://gist.github.com/karpathy/a4166c7fe253700972fcbc77e4ea32c5 using the OpenAI Gym and get the game running on your computer. Note that there is at least 1 change you will need to make to the code from version 2 to 3 of Python, and perhaps others depending on your libraries installed. Once running, check your code into Github and put a link to your repository (shared with me, look me up in GitHub using my Cestar email address) in your report. https://docs.github.com/en/get-started/quickstart/create-a-repo. Verify that only 1 CPU core is running your Python code and take a screenshot of your task manager showing this. 3. Determine the threshold % of games that the AI player will win that ends your test run. Instrument your code to measure the time taken until the AI player starts winning games, and the amount of RAM used by the program. Ensure that the same number of CPU cores (i.e. 1 core) is used for each test run. Document these values in your report. 4. Double the number of neurons in your code. Repeat your time and RAM usage measurements and document in your report. Update your Github repository by saving this change on a labelled branch. 5. Keep on increasing the number of neurons used until your system "breaks" (ie. becomes non-responsive, crashes, etc.) What value was this? What is the maximum number of neurons just before the breaking point for your system? Add your timing and RAM usage test results in your report. For each value tested, update your Github repository by saving the change on a labelled branch. 6. Now revert the number of neurons back to the default value of 200. Increase the learning rate from the default value. What effect does this have on the time taken and RAM usage? Decrease the learning rate from the default value. What effect does this have on the time taken and RAM usage? Collect at least 3 measurements in addition to the default value. Tabulate your results in your report. For each value tested, update your Github repository by saving this change on a labelled branch. Submit your report to this TurnItIn dropbox. A report template is provided in Moodle. karpathy.github.io (http://karpathy.github.io/2016/05/31/rl/)Deep Reinforcement Learning: Pong from Pixels Musings of a Computer Scientist.

Niko

امتیاز : 0/5

لیست پیشنهادها