yes everything runs fine in the colab
Send me your solution on private message I will have a look on it.
Hello, I have the same issue here.
The assignment is completed successfully in colab, but the grader gives the generic error:
There was a problem grading your submission. Check stderr for more details.
I read the thread here and:
- I’m not using additional imports
- Specified the loss with a string name
- Tried various optimizers as suggested but without success so far
What does the error say?
That’s the error:
“There was a problem grading your submission. Check stderr for more details.”
I don’t see anything else.
I am having the same issue.
Digit 0: IOU: 0.8796496269742625 Dice Score: 0.9359719113080295
Digit 1: IOU: 0.8996261118791311 Dice Score: 0.947161240049718
Digit 2: IOU: 0.8742637632851936 Dice Score: 0.9329143319217692
Digit 3: IOU: 0.841097081931364 Dice Score: 0.9136911792277992
Digit 4: IOU: 0.8344755649069197 Dice Score: 0.9097701608789327
Digit 5: IOU: 0.8510763192738233 Dice Score: 0.9195475199074453
Digit 6: IOU: 0.8597311501172835 Dice Score: 0.9245757378028161
Digit 7: IOU: 0.8807289771188579 Dice Score: 0.9365825569062818
Digit 8: IOU: 0.832350271288169 Dice Score: 0.9085056327172801
Digit 9: IOU: 0.8426724123961001 Dice Score: 0.9146198822180658
Grade is 85.95671279171106
You passed!
But once I upload the model for evaluation, I get a score of 0/100 and the error ‘There was a problem grading your submission. Check stderr for more details.’
The issue might be that the grader expects a certain hyperparameters, loss, optimizer, and metric…
If you are convinced that you model is built properly that the area I would look again.
Hi! I’m having the same problem and I already tried different losses and optimizers. Can you suggest me another solution to the problem?
2 Likes
Hi, I have also tried multiple models with different optimizers and passing the loss as a string and not a class but even though I get a good score on the notebook I always get the same generic error message on the grader “There was a problem grading your submission. Check stderr for more details.”. And yes, I’m convinced the model is built properly
It is also not working for me. I read all the comments in this thread and tried suggested solutions. It is impossible to debug on my side because error message is just “There was a problem grading your submission. Check stderr for more details.”
@Benny_Segers, @Sumeet_Mahesh, @gent.spah, @mariapedrososilva, @andres.gomez, @nat2: With so many people having trouble, it sounds suspiciously like something changed on the platform. We had a problem with another assignment due to a TF version mismatch now that Colab is using TF ver 2.11.
I haven’t had a chance to try this myself, but if anyone want’s to try the work-around we used for the other assignment, you can try downgrading the TF version in your assignment using these commands:
!pip install tensorflow==2.6.0
!pip install keras==2.6
Then re-run the model and create a new model.h5
I hope this helps! 
6 Likes
This worked for me! Thanks @Wendy
Thank you very much @Wendy !!, this worked for me as well.
It worked for me too! Thank you @Wendy.
Hi everyone! Sorry for the trouble with the lab and thank you for reporting! @Wendy is right and the notebook is now updated to freeze the TF and Keras versions to match the autograder. We’re now checking the other notebooks in this specialization for possible conflicts with the updated packages in Colab.
Great! Thank you for trying this. I’m glad it worked! Thanks, @chris.favila for the quick and thorough response.
Cannot downgrade TF and Keras in colab, looks like it is not allowed
!pip install tensorflow==2.6.0 --quiet
!pip install keras==2.6 --quiet
ERROR: Could not find a version that satisfies the requirement tensorflow==2.6.0 (from versions: 2.8.0rc0, 2.8.0rc1, 2.8.0, 2.8.1, 2.8.2, 2.8.3, 2.8.4, 2.9.0rc0, 2.9.0rc1, 2.9.0rc2, 2.9.0, 2.9.1, 2.9.2, 2.9.3, 2.10.0rc0, 2.10.0rc1, 2.10.0rc2, 2.10.0rc3, 2.10.0, 2.10.1, 2.11.0rc0, 2.11.0rc1, 2.11.0rc2, 2.11.0, 2.11.1, 2.12.0rc0, 2.12.0rc1, 2.12.0)
ERROR: No matching distribution found for tensorflow==2.6.0
so I did the assignment without downgrading.
My IOU Scores are as follows.
Digit 0: IOU: 0.909408824463932 Dice Score: 0.9525553802960445
Digit 1: IOU: 0.8212049215694128 Dice Score: 0.9018259415439579
Digit 2: IOU: 0.8428747085081641 Dice Score: 0.914739026605324
Digit 3: IOU: 0.8403732795671689 Dice Score: 0.9132639436764843
Digit 4: IOU: 0.8303080752201365 Dice Score: 0.9072877800861725
Digit 5: IOU: 0.8615209978870364 Dice Score: 0.9256097555331648
Digit 6: IOU: 0.8796465130037221 Dice Score: 0.9359701485552462
Digit 7: IOU: 0.8060806844551147 Dice Score: 0.892629760556135
Digit 8: IOU: 0.8700935704580222 Dice Score: 0.9305347969780137
Digit 9: IOU: 0.8350293525733281 Dice Score: 0.9100991778712817
Grade is 84.96540927706037
You passed!
- I had tried with different optimizers
- increased the number of epochs
- my model training resulted in , loss: 0.0110 - accuracy: 0.9957 - val_loss: 0.0221 - val_accuracy: 0.9929
The Predictions of the model looks good
When I uploaded the model.h5 file to the grader, I got 0/100, and the grader output was
There was a problem grading your submission. Check stderr for more details.
Is there anyone willing to look into my notebook privately?
Hi Nirban! Welcome to the community, and thank you for bringing this to our attention. We’ll work on this so you can submit your assignment successfully. In the meantime, you can send me the .ipynb file so we can also use it for testing. Thanks again!
Hello again Nirban! As a temporary fix, we updated the notebooks to install TF and Keras 2.8.0 instead of 2.6.0. Please update those lines in the pip install lines of your assignment, then resubmit the new model. The grader should accept that while we’re working on a more permanent solution. Thank you!
Downgrading TF and Keras to 2.8.0 worked, and my assignment scored 100/100. Thanks
2 Likes