Is anyone else having issues trying to run Lab1?
Kernel constantly dying during train the model (after [10]) . After a lost of restarts i have not yet got this to run successfully.
I suggest moving on to the next section for now and trying this lab again later. Sometimes kernel issues can be temporary. If it still doesn’t work after that, let us know, and we can inform the staff to look into a fix.
Also, please double-check that your code is correct and isn’t causing the kernel to crash. Sometimes errors in the code can cause kernel issues, so a quick review might help!
Yes, Ill do that. I deleted cache/cookies/history again (in chrome browser running on win10 x64) and started from scratch. Still get a fail;
1 Like
its not my code - it is Lab1. I havent made any changes to it. However i will try it with batch size = 16 , and see if that helps.
1 Like
Sometimes trying with different browser can be helpful, too!
Sure! Let me know if that resolves it, and if not, we’ll explore other options!
Can I first confirm if the specialisation you selected is correct, because I saw you working on tensorflow Developer professional certificate course and you posted your query in NLP specialisation
This issue has been reported, it is because of tf and keras version discrepancy.
It’s ungraded lab right?
ys ungraded lab. the name of the course is: Natural Language Processing in TensorFlow. week3 lab 1. I am abandoing this. It got further with batch=16, but then the jupyter lab connection time out after 2 hours plus and i cant jst sit and watch it for that long
. I am abandoing this. I got further with batch=16. I got to 9 epochs. But that took 2 hours, and then the jupyter lab connection timed out after 2 hours plus .
check the keras version lab is working on. For now I have reported your issue to the staff.
Thank you for your understanding and patience.
Regards
DP
Hi Cormac. Sorry for the trouble with the lab. Unfortunately, I can’t reproduce the issue. It’s working fine on my end. I assume you’re doing the lab without any modifications so it should also run fine. I also assume you’re working inside Coursera Labs, and not on your own local installation of Tensorflow. If it’s still slow on your next attempt, (one epoch should take just around 30 to 40 seconds) please send me your lab ID. You can click the question mark icon on the upper right of the lab, and find the lab ID at the bottom. Thanks.
1 Like