C5W4A1 Transformer Architecture with TensorFlow submission error

[htsd18]

Please help me get the certificate!!!

UNQc7 and UNQc8 are not functioning (kindly fix the Autograder)

[htsd18]

[sjfischer]

@htsd18, this might be an Autograder issue, but just to be sure: did you correctly pass all tests before submission without fail? If yes, can you resubmit and then share the Autograder output in this threat. I can then submit an official issue to the Coursera team to get this fixed. So far, no Autograder issues have been reported for Course 5.

Please bare with us,
thanks Stephanus

fyr @Mubsi

[htsd18]

[Week4] Transformer Network Decoder_test/Transformer_test

[image] and [image] How to solve this?

{mentor edit: hacks removed - no longer necessary}

[htsd18]

[Week4] Transformer Network Decoder_test/Transformer_test

[image] {mentor edit: hacks removed - no longer necessary}

[innav1981]

Same problem here. Commenting the failing asserts doesn’t help to pass the overall assignment.

[htsd18]

{mentor edit: hacks removed - no longer necessary}

[htsd18]

{mentor edit: hacks removed - no longer necessary}

[jktubs]

@htsd18
in the decoderLayer are 2 mha layers. The first one gets only x as the 3 inputs, the second one, e.g. for me:

{mentor edit: code removed}

[jktubs]

{mentor edit: hacks removed - no longer necessary}

[quake1st]

As of today Jul/16, I didn’t apply the above hack, and was able to get grade 100/100. So the following worked for me:

self.mha2(out1, enc_output, enc_output, …)

I guess there are some changes in the backend that resolved the problem which this hack workarounds. Just FYI.

[TMosh]

The hack workarounds are not necessary.
The exercise has been updated in the last two months.

[Juan_Ricardo_Escobar]

DeepLearning.AI
Instructor,

I hope this message finds you well. I am writing to bring to your attention an issue regarding the grading of my final programming assessment in the Sequence Models course at DeepLearning.AI.

According to my records, I successfully passed all the tests of the final programming assessment. I rigorously ran all the codes using my programming skills, and the results indicated that I passed each test without any errors. However, upon submitting the assignment, I noticed a discrepancy in the grades I received. Some grades were marked as passed, while others were marked as failed.

Given that all the tests were passed when I ran them locally, I believe there might be a mistake or an oversight in the grading process. I kindly request your assistance in resolving this issue and ensuring that my grades accurately reflect my performance in the assessment.

I understand that grading can sometimes be a complex process, and errors can occur. I am more than willing to provide any additional information or clarification that may be required to rectify this situation promptly. It is important to me to have an accurate evaluation of my work and efforts in the course.

I appreciate your attention to this matter and I kindly request that you look into the grading discrepancy as soon as possible. I look forward to your prompt response and a resolution to this issue.

Thank you for your understanding and assistance.

Sincerely,
Juan Ricardo