Dot vs matmul vs @

I noticed in the backwards propagation algorithm in the labs and assignments, it uses np.dot over np.matmul (and @ operator). Was wondering why? I didn’t see this mentioned in the course anywhere.

They are equivalent in many situations.

Personally I avoid the @ operator, because it’s overloaded and you don’t know exactly what function it’s going to implement. I find it confusing.

1 Like

Thanks! The choice between dot and matmul is just stylistic in these cases? Performance differences, etc?

Sorry, i have not studied that.

this link should help you understand difference between np.dot and np.matmul