Viewing profile — lelag

This is all fairly obvious, no? At first, write clean code with functions and don’t obsess over call overhead. Once it works, profile, then optimize where it actually matters. Prem…

Indeed. I've tried to run it locally this but couldn't get it running on my measly gaming-spec workstation. It's seems you need lot's of ram and vram. Reading the issues on github[…

Miracle Max gave us a clear definition if I recall, you die when you are "all dead", as long as you are mostly dead, you are slightly alive... I'll let myself out now.

In his last blog post, Sam Altman also revealed how much power the average chatgpt query uses, and it's in the same ballpark. > People are often curious about how much energy a Cha…

Damn, I would not have guessed that Men In Black was actually a documentary...

Maybe they want to compile the Apollo Guidance Computer source code... https://www.softwareheritage.org/wp-content/uploads/2019/07/...

The metabase "backend" is written in clojure. The web frontend is written in TypeScript/React.

Interesting point about PC going digital-only as Nintendo is a fascinating counter-example. While they offer digital downloads on the eShop, their pricing actively discourages it. …

If I understand it correctly, that's a valid concern but the way structured generation library like outlines[1] work is that they can generate multiple variants of the inference (w…

I can’t say for certain, but I’d guess that writing without the letter “e” is slightly more difficult in French than in English. For one, “e” is a bit more common in French (around…

I was going to point that out. What I will add is that constrained generation is supported by the major inference engine like llama.cpp, vllm and the likes, so what you are describ…

More like 2000 tsar bombas if your energy release calculation is correct.

OSS model do not have to be local models, and it's not just about privacy, imo. DeepSeek R1 hosting is out of reach for most, but it being open is a game changer if you are a build…

Aren't we already living in the Biff got rich timeline?

Thank for the explanation. I guess it makes some sense, considering many people with no nlp background are using those models now…

Yes, weird that they didn't use that term for this project.

If that's an issue, there's a workaround using structure generation to force it to output a token after some threshold and force it to write the final answer. It's a method used to…

> And where does the nuclear fuel come from? Russia. Not true at all. Russia is producing 5% of the world Uranium, and they probably use quite a lot of that domestically given they…

The problem with the original Kinect (v1) is that good tracking software for it was never really written. Most application that support it, just use the original Microsoft SDK to d…

Saying this is similar to Moshi is like saying GPT2 is similar to GPT4. You can't have any sort of conversation longer than 30s with moshi before it goes banana. You can talk to th…

Oh, right. I misunderstood.

I'm happy to see you repaired your keyboard.

The SMPL-x body model, a standard in this academic field does model fingers https://smpl-x.is.tue.mpg.de/ The issue is that there are much less dataset available for it than for th…

Some more thoughts about training a manipulation model: I would add that synthetic data might be key to making it happen. One issue is that most video is not shot in first person, …