What should I do tonight?

[gstvtrp]

I haven't left my apartment in about 28 hours.

Completed part 1 of this today: http://cs231n.github.io/assignments2018/assignment2/

Don't feel like doing any more. Tomorrow I'll attempt batch norm.

Other things I need to do:
Haven't eaten dinner yet.
Apartment could use some cleaning/organizing
Need to get quarters for laundry

 

[Weightcel]

Fap or clean.

 

[Deleted member 60]

go ER

 

[idkwattodowithlife]

Clean, then eat/relax/watch something or browse Incels.me for a while. Back to working on your assignment. Watch something or go back to browsing incels.is for a while, then go to sleep.

 

[gstvtrp]

Clean, then eat/relax/watch something or browse Incels.me for a while. Back to working on your assignment. Watch something or go back to browsing incels.is for a while, then go to sleep.

I cleaned up a little and ate. Now I'm here lol.

 

[idkwattodowithlife]

I cleaned up a little and ate. Now I'm here lol.

Lol'ed.

 

[blackcel]

bruh Bruh BRUH

Have you tried anything on kaggle yet?

 

[gstvtrp]

bruh Bruh BRUH

Have you tried anything on kaggle yet?

nah, I doubt I could win those competitions.

Besides, it seems like Random Forests is the most effective method on those datasets. Not really interested in random forests.

 

[TheGoat]

Beat your meat so hard that yur penis reports you for abuse

 

[blackcel]

nah, I doubt I could win those competitions.

Besides, it seems like Random Forests is the most effective method on those datasets. Not really interested in random forests.

Random forests is a very versatile algorithm that is hard to fuck up, which is why so many people seem to use it. I wouldn't say random forests are best for most datasets on kaggle, but what is best depends on who you ask.

I like neural networks a lot as well but they are extremely sensitive to the data that you feed it. If your pre processing is shit then the network will return shit. Neural networks are nothing more than convex optimizations hidden behind layers upon layers of convoluted bullshit. If a solution exists within the given parameters then it will find it. Problem is that you have to feed it the right parameters and train it correctly, which is an art on its own.

Anyway I mentioned kaggle because it provides you with a lot of datasets to just test random shit on. It doesn't matter if you win or not, it is just a learning exercise. You can also put that shit on your resume to show future employers that you kinda know what you are doing.

 

[LebaneseCel]

Rope

 

[Anger]

LDAR

 

[Deleted member 3291]

cool Im going to start reading some more papers when I get more of the math to my understanding.

Also take a look into computer cryptography/encryption it’s a pretty broad field and hard to generalize

 

[GreenBull]

Exercise.

 

[PM_ME_STRIPPERS]

drink