It's still better than providing a code full of bugs at release date.

Deep learning movie: Overfitting strikes back.

Regular developers would think of it as a JSON 😂

Yeah but on test data*
It ain't gonna happen 😂

I would be worried 😂

Machine learning enthusiasts really like using deep learning neural networks for simple tasks 😂

Average data scientist be like

So true 😂

Those who tried AWS understand the struggle 😂

I guess it's overfitting 😂

Exactly! I have been using many pretranied models throughout my career, especially for NLP task and it's completely true that i am more angry when that models make mistake or if they are incomplete, compared to modela i trained myself.

All of us felt like this sometimes 😂

It's not a bug, it's a feature 😂

Free wifi 😂😂

Do you use Jupiter notebooks 😂😂

Math and machine learning can't exist without each other 😂

Old Statistics joke

Just give me more data 😂

My conventional neural network performance 😂

I guess I should have trained it on a few more epochs