I did a number of assignments in C1 and C2 (for example, C2W1) 2 days ago and successfully submitted them. I hit the “save” button inside the DeepLearning.AI Jupyter instance and it updated the modified time for the .ipynb file.
Today I went back to reference some of that work, in completing a new assignment, but I found that it gave me a brand new instance with the default, incomplete assignment files, with modification times of several weeks ago. Is this expected behavior? It’s not how the Coursera courses worked when I took them.
If possible, I’d like to be able to download the assignments I submitted. If this isn’t a bug, the course should certainly warn you that your finished assignments are wiped within X hours.
Just on general principles, it’s always a good idea to save your work “early and often”. The issue is that on the new DLAI learning platform whenever they publish a fixed or modified version of an assignment, there is no way for them to preserve any of your existing work: it just overwrites the file. On Coursera, it will move your work to a new file name with the date and time interpolated into the file name, but that capability does not yet exist on the new platform.
This is early days both for the new platform and for any of the courses that are newly ported to it. Meaning that there is a lot of activity with bugs being reported and fixed, so my guess is that you just got unlucky and they pushed updated versions of those assignments since you last worked on them. I’ve had the kernels freeze pretty frequently and that can also cause you to lose any changes since the last time you did a “save”. So “save early, save often” (meaning both save on the platform and download the current state of the ipynb file to your local computer) is the way to go.
Thanks for the advice; I appreciate it. I am still looking to hear from DLAI support on the matter, since I’d like to hear from them if this is an intended customer experience. It’d certainly be a factor in choosing to continue my subscription.
Yes, they are, but they generally are pretty busy and don’t always read things anywhere close to realtime and are not guaranteed to notice any particular thread. I can “tag” them, now that I know you are expecting a response from someone on the staff.
Most of the active support is from the mentors, who are fellow student volunteers. Meaning that we do not get paid to do this and have no employment relationship with DeepLearning.AI. I do not know the TF courses, so I generally don’t respond to questions about them. But I noticed this one and thought it was more a question about the new platform, about which I do have some recent experience.
Hi, and thank you for reporting this. I’ll forward this to the right department and see if they can recover your notebooks. Will also clarify the workspace behavior you mentioned. I might need some details though. Will send you a direct message. Please check your inbox.