marq.ts @rollitomaki Twitter Profile

marq.ts

@rollitomaki

nn.Sequential(humor, malo)

2KPosts 71Followers 204Following

Similar User

@cringe_salad

@satelec_etsit

@paulaesteban47

@peanutsfreezone

@dhdezcorral

@TitoSamu13

@Olgyyybf

@angelit0_7

@Hertz_io

@nightmxller

@WillyDontLie

@RazerRedFox1

@pulpotorro

@lwjist

Pinned

marq.ts

@rollitomaki

23 May 2022

aaaa

marq.ts Reposted

Jose Mª Echarte

@j_echarte

18 Nov

Spanish Ramen.

marq.ts Reposted

marq.ts

@rollitomaki

17 Nov

nice so installing and maintaining torch has just become more fucking difficult

We are announcing that PyTorch will stop publishing Anaconda packages on PyTorch’s official anaconda channels. For more information, please refer to the following post on dev-discuss: dev-discuss.pytorch.org/t/pytorch-depr…

marq.ts

@rollitomaki

10 Nov

????? wow

Shohei Taniguchi

@ishohei220

6 Nov

Our NeurIPS paper is published on arXiv. In this paper, we propose a new optimizer ADOPT, which converges better than Adam in both theory and practice. You can use ADOPT by just replacing one line in your code. arxiv.org/abs/2411.02853

marq.ts Reposted

kalomaze

@kalomaze

7 Nov

arxiv.org/abs/2410.21228…

marq.ts Reposted

ily⚡️

@0xIlyy

7 Nov

Mfs will turn on a LED and then say you can’t make it in CS without hardware😭

brolic boi

@bubblebabyboi

7 Nov

We are going to make it

marq.ts Reposted

jack morris

@jxmnop

6 Nov

brutal rejection from NeurIPS

marq.ts Reposted

Ethan 🇦🇺

@torchcompiled

5 Nov

marq.ts Reposted

konrad

@konradgajdus

3 Nov

Hardware level CNN in system verilog Achieving good performance and low latency in simulation Fully synthesizable

marq.ts Reposted

laurence

@functi0nZer0

1 Nov

marq.ts Reposted

@bluecow 🐮(schizo)

@BLUECOW009

28 Oct

colab.research.google.com/drive/1Dl2ETy-…

marq.ts

@rollitomaki

26 Oct

my thirteenth reason

marq.ts Reposted

angel 𓆩𓆪 virtual

@cyberdaemonum

23 Oct

😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭😭

marq.ts Reposted

Alperen

@AlperenAkkuncu

22 Oct

2-pin 2.54mm headers are great test points for oscilloscope probe connections. However, it's a bit unsafe because you can easily reverse the probe accidentally, which connects the ground spring incorrectly. You might damage your oscilloscope input 😬 It's a great low inductance…

marq.ts

@rollitomaki

22 Oct

jaja osea q ahora resulta que attention is not all you need de hecho parece q ni lo necesitas ???

Sebastian Raschka

@rasbt

22 Oct

"What Matters In Transformers?" is an interesting paper (arxiv.org/abs/2406.15786) that finds you can actually remove half of the attention layers in LLMs like Llama without noticeably reducing modeling performance. The concept is relatively simple. The authors delete attention…

marq.ts Reposted

.

@marianawismann

20 Oct

Acabo de leer que “los 20s se sienten como si estuvieras yendo tarde a algo importante pero no sabes a qué” y wey no mms I felt that.

marq.ts Reposted

✧

@northstardoll

11 Oct

marq.ts Reposted

ً

@sonoreid

18 Oct

being called smart because you have a variety of information on different subjects but in reality it’s all surface level intelligence and you don’t feel like you’re really good at anything