I got this error log
now I don’t know if python3.8 will work in my MacBook 10.15
I have to uninstall the python3.11.3
can someone answer these doubts @chris.favila
Elirod
can the conversion fail if my chosen path differs from the assignment mentioned??
yep!
you have to pay attention on this
Hi @Deepti_Prasad ! May be this will help you:
I did not try it myself though, since I use Windows and Ubuntu on PC, not Mac. But you saw how I handled the issue on my system in another thread in this forum.
By the way, ask yourself: why do I need “conda” for this assignment? My answer is simple: you do not need “conda”. Keep things simple: just use “pip”, if you can. At least this is a way I approach this task.
1 Like
Hello @Deepti_Prasad ,
I see it’s 3.11 version.
- Remove it :
Command -
sudo rm -rf /Library/Frameworks/Python.framework/Versions/3.11
sudo rm -rf /usr/local/bin/python3.11
Then download fresh Python 3.8 .
- I haven’t done this specialization, so I don’t know if only Jupyter is mandatory. If it’s not, you can use Pycharm IDE. There you can select Particular versions of Python as the Python compiler.
With regards,
Nilosree Sengupta
1 Like
use terminal to remove it
1 Like
Hey nilosree,
I don’t want to remove my current version of python 3.11 and use only 3.8 for the assignment I am doing and I have found the solution by @James_J_Johnson. Yes Jupyter is necessary for this course. Thank you for your inputs.
Regards
DP
Thank you James, even I was stating the same, conda is not required. Thank you again for replying here and on your thread promptly to my issue. I will look into the link you sent, your thread helped me for this course assignment week 3. But what I noticed people who have tensorflow version 2.1 where able to do this assignment, probably the latest version is not compatible with the converter. Keeping 
Regards
DP
I am glad to help. Tensorflow 2.1 is really old. The most recent is TF 2.13. But I use TF 2.8.2 on Windows and TF 2.9.1 on Ubuntu (WSL2), and both support GPU. I had to downgrade from TF 2.12, since it had some bugs with some of my models. My versions of TF are compatible with whatever converter that I install with “pip” in Python (3.10.8 on Windows and 3.10.11 on Ubuntu). The problem is different: the grader is not updated to handle these more recent versions of TF or converter.
Hello @Deepti_Prasad ,
Oh I see.
Im glad you got the solution from James.
Happy learning!!
With regards,
Nilosree Sengupta
Hi Deepti. As mentioned in the note in the instructions for that assignment, there is an ongoing issue with the grader. Please monitor this topic for updates. We’ll try to have a workaround by end of day. Thanks.
Hey Chris,
Always happy to hear from you 
So based on the topic you told me to check, tells now I can use Coursera Lab environment with the TF 2.2 version for doing the TF Data and Deployment course assignments.
But our issue was not that. I was okay to work on Jupyter notebook. The issue was we all are using upgraded version of tensorflow, like I am using 2.11 and that is not compatible with the exercise conversion to json file. So we wanted to know when the assignment will be compatible or work with the current version of tensorflow or any upgraded version of tensorflow other 2.2 as most of us have tensorflow version beyond 2.8.
Sorry still bothering you 
Regards
DP
Hi Deepti. I think the topic addresses your concern. Unfortunately, we can’t give an exact timeline for updating the grader, but it is on our To-Dos. For now, please use the Coursera lab for the Week 3 assignment. Please read through the instructions, and also the markdown in the Coursera version of the assignment. That should explain why it’s done this way. What you will do in the Coursera lab is still applicable to newer TF versions. Thus, I don’t think you should wait for the autograder update to do the assignment. That’s because you will write the same commands anyway. The problem is just the compatibility with the current autograder because TF updates/deprecates internal model parameters between versions. You won’t see that in the code that you have to write in this notebook. After you get a passing grade, you can try exporting the notebook and the dataset on your own environment/Colab to see that the commands still work. You will save the model, and convert it to JS the same way. The only difference is you can’t submit that to the current version of the autograder. I hope this is okay for the time being.
1 Like
Chris
Yes the Coursera lab environment helped with the grader issue, and I want to inform and thank you for all the response and work QA team does. Thank you again.
Regards
DP
1 Like
James they created Coursera Lab environment which didn’t require me to do all the task you had to do for completing this assignment. Thank you again for all the response and support
Regards
DP
Hi how can I start this
Hi Varsha,
your question is unclear, please elaborate what you want to start?
Regards
DP
I can’t understand this …how can I do this what benifit u get
Varsha please explain when you are stating this?
What is this in your statement??
are you asking about TensorFlow Data and Deployment Course?
or you are asking about this particular weeks’ assignment?
or you are asking what are benefits of doing TF Data and Deployment Course?
Please be clear…this is very confusing word here in this pretext
Regards
DP