C3-Day5-VertexAI Assignment

Course Q&A Machine Learning Engineering for Production(MLOps) Machine Learning Modeling Pipelines in Production
23

Ok I managed to complete the assignment.

One should create user managed notebook rather than instance. And I used a zone from london/europe.

Part of the assignment requires user to train the model and that takes about 20mins. And you have to run it twice.

The training might seem stuck as it takes very long. So the only way to check is to click on Check My Progress. If cleared, you can stop the training and move on.

Grades will be updated in Coursera page when each step of Check My Progress is cleared.

All the best.

1 Like
24

Mine worked with this, thank you so much. And at the end, you totally don’t have to wait 20 minutes for it to train - I got my certificate right after I entered the training command, so I just exited out lol

25

for now, i cannot login with my account, can you check it for me pls?

26

Machine type e2-standard-4 (default selected) worked for me, presumably that’s a little more beefy than e2-standard-2.
Another issue that happened for me is the default Debian image does not seem to have graphviz installed, and the tf.keras.utils.plot_model step failed for me. Running a cell with !sudo apt install -y graphviz solved that.

27

code block of train and evaluation is empty! where do you find the code?

28

code block of training and eva is empty, where is the code?

training and eva

29

Anyone getting this error? I can’t even access the assignment instructions.
image

30

Hi Michael. Thank you for reporting, and we’ve escalated it to our partners. Kindly monitor this topic for updates.