Content pfp
Content
@
0 reply
0 recast
0 reaction

Vitalik Buterin pfp
Vitalik Buterin
@vitalik.eth
The differences between the APIs of numpy, cupy and torch are so fascinating.... ``` >>> import torch as np >>> np.arange(10) tensor([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) >>> a = np.zeros(20) >>> a[19:-1:-2] = np.arange(10) ``` Torch doesn't let you have ranges that go backwards 🤣 Would love more consistency
43 replies
641 recasts
2642 reactions

Kevin pfp
Kevin
@typedarray.eth
Highly recommend trying Polars (https://docs.pola.rs) if you haven't. The APIs are very thoughtfully designed. May not be suitable for your use case if you need GPU acceleration though (https://pola.rs/posts/polars-on-gpu).
0 reply
22 recasts
403 reactions

yesyes pfp
yesyes
@yesyes
Are you trying out data science/ML?
0 reply
0 recast
1 reaction

Ajit Tripathi pfp
Ajit Tripathi
@chainyoda
Ser is pivoting to ai 💕
0 reply
0 recast
0 reaction

kevin mfer 🎩 pfp
kevin mfer 🎩
@kevinmfer
idk any coding but i felt that 😂😂😂
2 replies
1 recast
3 reactions

Jason Cheuk pfp
Jason Cheuk
@babycheuk
When NumPy, CuPy, and PyTorch meet at the library, it's like a superhero showdown! NumPy's the classic hero, CuPy's the edgy anti-hero, and PyTorch? Well, PyTorch moonwalks uphill while sipping gradients. 🕺🔥
0 reply
0 recast
1 reaction

Mark 🎩 pfp
Mark 🎩
@web3withmark
Yo @dwr.eth @v can we get a code highlighter on this app? Let’s make @vitalik.eth tweet beautiful 😍
0 reply
0 recast
1 reaction

Marcel🔵 pfp
Marcel🔵
@4cademy.eth
'import torch as np' hits different xD
0 reply
0 recast
1 reaction

jenny.degen 🎩🟣 pfp
jenny.degen 🎩🟣
@cryptojenny
So true!
0 reply
0 recast
1 reaction

Flacko0x pfp
Flacko0x
@ashotshaitan
Думаете он разбирается?
0 reply
0 recast
1 reaction

0xqeew  pfp
0xqeew
@0xqeew
If you don't understand this, use me as a like button
0 reply
0 recast
1 reaction

Anemale🎩↑🔵 pfp
Anemale🎩↑🔵
@anemale.eth
I like it, but I don't understand anything at all.
0 reply
0 recast
1 reaction

zpuheng pfp
zpuheng
@zpuheng
200 $DGGEN
0 reply
0 recast
0 reaction

freely pfp
freely
@freely
very like it
0 reply
0 recast
0 reaction

DigiJourney pfp
DigiJourney
@svosdamage
It's definitely intriguing how each library has its own quirks! Numpy and CuPy are almost seamless twins, but Torch does take a different path. It's a blend of flexibility and complexity. Consistency would be great, but exploring these differences broadens our understanding and adaptability across frameworks!
0 reply
0 recast
0 reaction

melniker pfp
melniker
@telaik
vitalik vitalik
0 reply
0 recast
0 reaction

Alan pfp
Alan
@give
it is beautiful
0 reply
0 recast
0 reaction

木木 pfp
木木
@numu001
I like it if though i don't understand it
0 reply
0 recast
0 reaction

Frontcaster pfp
Frontcaster
@frontcaster
I‘m very happy I find $tensor
0 reply
0 recast
0 reaction

jjmm pfp
jjmm
@wenwen666888
interesting point
0 reply
0 recast
0 reaction