Hello, this course is so good that, after six years, I am repeating it. The first time was on Coursera, and I was using MATLAB. Now it’s using Jupyter Notebooks. I would like to download the Jupyter Notebooks and run them on my computer using Colab from Google, but I can’t. Any help out there on running the Jupyter Notebooks locally?
Hi @snowflake1 ,
If you enrol under the audit option, you will have access of the first module material free of charge. But all subsequent graded assignments/labs are locked behind a paywall which requires active subscription for access.
Kic, thanks. I am enrolled as Pro, so should have access. Where can I download the module material exactly? Regards
Hi @snowflake1 ,
The Deeplearning.ai platform operates slightly differently from Coursera, and you will have to download the files individually.
Open the Jupyter notebook, on the menu bar, and click:
File → Open, click on the check box to select the lab assignment file with a green icon.
click on the Shutdown tab to shutdown the Kernel, a different menu with Download option appears. Click the Download
Yes, they haven’t (yet anyway) made downloading as convenient on the DLAI Learning Platform as it is on Coursera. There is also a way to package up all the files and download en masse, but you have to be comfortable with linux shell commands. Please see this thread for instructions.
In terms of running things locally, you are on your own. There are too many possible variations for it to be reasonable for there to be official instructions. There are some threads on the forum that will get you started down that long and winding road, but the point is that if you’re going to play that kind of game you need to build up your own problem solving skills. Stack Exchange is your friend. 
Or here’s another such thread.
Sorry! I missed the fact that you asked about using Colab. That is actually somewhat easier than running things locally on your own computer, although you can still have “versionitis” issues. Colab also supports Jupyter Notebooks, but the way the file system works is a little different: you need to use your Google Drive instead of a direct Docker image. Here’s a thread from mentor Saif which shows how you need to modify the notebook code in order to handle the different file access.
1 Like
One other point worth making about Colab is that it is not safe to upload the notebook to Coursera or the DeepLearning.AI learning platform after you have run it on Colab, even if you revert the file management code. The problem is that there is a lot of invisible JSON code in the notebook and Colab modifies that “metadata” in ways that are incompatible with the metadata that the Coursera version of the courses uses. I haven’t tried this experiment on the DeepLearning platform yet, but based on the bad experiences with Coursera it would advise against it.
You could “copy/paste” your python code from the “YOUR CODE HERE” sections of the notebook back to your notebook on the course platform, but for the full notebook file please consider it a “one way” path from the course platform to Colab.
1 Like
Kin, thank you so much. I confirm that I have been able to download all the files and have been able to run thw first ones in my 2 machines, Linux and Windows. Regards
Thanks Paul,
The thread is very good, and touches what I feel will be better an increasing paint point over the coming weeks. Much Appreciated!
Paul, thanks again. I have been able to run the files locally in a Windows and Linux machine running VS Code with Python, Junyent and Conda extensions. I prefer these as Colab is on the cloud. But now I know that I could try to run them on the cloud as well. Thanks again
Paul, you are really great at this. I decided to be hands off Colab after what you wrote Regards