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

Hey everyone, Just to clarify, the assignment has been updated on 22 May. You don’t need any hacks to accomplish it since then :slight_smile: If you don’t notice changes in your assignment, try refreshing your workspace .

Same question here. I was following the TF transformer guideline.

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

I have the same issue. Problem is something else in my opinion. I got the same output as you.

My DecoderLayer part works fine, but the decoder fails. This is my decoder call function coding part: {mentor edit: code removed}

Mine is same as yours, Decoder call still fails. Here is my Decoder Layer, I have commented out two assertions and it passed the tests: {mentor edit: code removed}

[image] GangYao: self.mha2(out1,enc_output, enc_output, padding_mask, return_attention_scores=True) assert np.allclose(outd[1, 1], [-0.34560338, -0.8762897, -0.4767484, 1.6986415]), “Wrong values in outd” Still the same error.

When I have done it Decoder_test fails as follows: assert np.allclose(outd[1, 1], [-0.34560338, -0.8762897, -0.4767484, 1.6986415]), “Wrong values in outd”

[Week4] Transformer Network Decoder_test/Transformer_test

Course Q&A Deep Learning Specialization Sequence Models

GangYao May 9, 2021, 12:34pm 5

{mentor edit: hacks removed - no longer necessary}

C5W4A1 Transformer Architecture with TensorFlow submission error

Topic		Replies	Views
C5_W4_A1_Transformer_Subclass_v1 : UNQ_C8 Sequence Models coursera-platform	8	792	November 5, 2021
Correct code but still giving error- unable to figure out Sequence Models coursera-platform	4	557	January 12, 2022
C5W4A1 Transformer Architecture with TensorFlow submission error Sequence Models coursera-platform	12	1053	June 2, 2023
[Week 4] Assignment 1 >> Decoder class error Sequence Models coursera-platform	65	3116	February 27, 2025
Ex 7 decoder wrong output Sequence Models coursera-platform	9	1360	July 15, 2022

[Week4] Transformer Network Decoder_test/Transformer_test

Related topics