p1esk

8 ms·

211.
by p1esk 10mo ago
It is right. The shuffling is from CPU memory to GPU memory, and from GPU memory to GPU. If you don’t have enough memory you can’t run the model.
212.
by p1esk 10mo ago
You’re describing this like if you actually knew what’s going on in these models. In reality it’s just a guess and not a very convincing one.
213.
by p1esk 10mo ago
Have you been living under a rock?
214.
by p1esk 10mo ago
I’m an ML researcher currently working on model compression for AI hardware accelerators. Mostly developing and testing quantization algorithms, and hw/sw codesign.
215.
by p1esk 10mo ago
I thought everyone switched to cursor by now, why do people still use vscode?
216.
by p1esk 10mo ago
Unless AGI is just a bunch of CS algorithms.
217.
by p1esk 10mo ago
I’m not defending this guy, but many fathers leave because they don’t want to have family. It could have nothing to do with any of his other plans.
218.
by p1esk 10mo ago
in the academic world 20 papers with 50 citations is worth more than one paper with 1000 It depends. If your goal is to get a job at OpenAI or DeepMind, one famous paper might be better.
219.
by p1esk 11mo ago
Do they really think an army of robot dogs is going to protect them? Why would anyone think otherwise?
220.
by p1esk 11mo ago
Until google puts in a lot of resources into training a scaled up version of this architecture If Google is not willing to scale it up, then why would anyone else?
221.
by p1esk 11mo ago
"We have demonstrated that it is impossible to describe all aspects of physical reality using a computational theory of quantum gravity," says Dr. Faizal. "Therefore, no physically complete and consistent theory of everything
222.
by p1esk 11mo ago
Pretty sure it was always like this
223.
by p1esk 11mo ago
These engineers make enough money to comfortably retire by the time they are replaced with AI.
224.
by p1esk 11mo ago
I’m not blind and AI has been great for me too.
225.
by p1esk 11mo ago
Did you mean: “if your AI can’t write a paper that passes an AI detector, are you any good as an AI researcher?”
226.
by p1esk 11mo ago
Unstructured sparsity cannot be implemented in hardware efficiently if you still want to do matrix multiplication. If you don’t want to do matrix multiplication you first need to come up with new algorithms, tested in software. This reminds
227.
by p1esk 11mo ago
This seems like poetry
228.
by p1esk 11mo ago
The big thing is that sparse models allow you to train models with significantly larger dimensionality, blowing up the dimensions several orders of magnitudes. Do you have any evidence to support this statement? Or are you imagining some
229.
by p1esk 11mo ago
Yes, we know that large dense layers work better than small dense layers (up to a point). We also know how to train large dense models and then prune them. But we don’t know how to train large sparse models to be better than large dense mod
230.
by p1esk 11mo ago
I’m not sure why you’re talking about efficiency when the question is “do sparse models work better than dense models?” The answer is no, they don’t. Even the old LTH paper you cited trains a dense model and then tries to prune it without t
231.
by p1esk 11mo ago
“Useful” does not mean “better”. It just means “we could not do dense”. All modern state of the art models use dense layers (both weight and inputs). Quantization is also used to make models smaller and faster, but never better in terms of
232.
by p1esk 11mo ago
I could not find any evidence that sparse models work better than dense models.
233.
by p1esk 11mo ago
Deep Learning models would work way better with much higher dimensional sparse vectors Citations?
234.
by p1esk 11mo ago
Yes, if you have a billion dollars then in terms of wealth Pichai is closer to you than to Bezos. But if you’re a typical HN reader (level 4 or 5), the difference between you and Pichai is pretty much infinite, while Pichai and Bezos are al
235.
by p1esk 11mo ago
“The difference between Pichai and Bezos dwarfs the difference between Pichai and me” I don’t understand: Pichai is a billionaire.
236.
by p1esk 11mo ago
If you’re familiar with Marxism you don’t need this article, but most people are not.
237.
by p1esk 11mo ago
I think these groupings are meaningful: there are pretty clear social boundaries between these levels.
238.
by p1esk 11mo ago
Why?
239.
by p1esk 11mo ago
Two reasons I decided not to go with tankless: 1. Can’t have recirculation pump to supply hot water instantly throughout the whole house. 2. Significantly higher energy bill compared to heat pump. An additional bonus is the heat pump cools
240.
by p1esk 11mo ago
I thought LG G5 is the best OLED TV in 2025

More ›