Week 2 Assignment 2: v1 model does better than it should 😉

Course Q&A Deep Learning Specialization Sequence Models
1

The second assignment of the word embeddings week uses the sentence ‘not feeling happy’ to show how the v2 LTSM model is better than the simple v1 softmax network. However in my case, both networks correctly associated the sentence with the correct emoji (:disappointed:). Some of the other v1 inputs were clearly not respecting order, though, such as ‘not happy’ - but my v2 network also didn’t get ‘not happy’ correct: it returns

not happy :smile:

I didn’t dig deeper and this didn’t prevent me completing, I’m just posting as an FYI.

I’ve attached samples from my v1 emojifier.

Screenshot 2023-05-15 at 11.54.22 AM

2

Thanks for posting your results.