DLS Sequence course - #Week2 Assignment2 - Test result problem

Hello! I was doing my assignment about emoji, and i found something strange.
I did well on code and get 100 accuracy at training, and 85% on test set.
But at last part, ’ Run the cell below to see the mislabelled examples:’
Here i don’t think it makes good accuracy. What is the problem? Maybe on local environment? Or data problem?
Can there be a One hot encoding problem? But i don’t touch anything about OHE


Hi @NothingON

The text mentions “mislabelled examples,” which means the output is specifically showing only the incorrectly predicted samples for you to review. This doesn’t reflect the overall accuracy but rather highlights the cases where the model made errors.

Since your test accuracy is 85%, it’s normal to have some misclassified examples, these are just the 15% of predictions that were incorrect. It’s not an issue with your code or one-hot encoding.

Hope it helps! feel free to ask if you need further assistance.

3 Likes

I appreciate you. It is cause of my English skill. My bad.
Thx a lot!

No problem at all :grin: You’re welcome!