Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
eurekin
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
32 ms
·
451.
▲
by
eurekin
3y ago
This is huge news... Has anybody seen any discussions around that topic?
452.
▲
by
eurekin
3y ago
I think this is relevant: https://www.whitehouse.gov/briefing-room/presidential-action...
453.
▲
by
eurekin
3y ago
Are there any external hardware encoders/decoders?
454.
▲
by
eurekin
3y ago
Didn't see batching taken into equation, might skew a bit
455.
▲
by
eurekin
3y ago
Interesting! I'd have one use case for classification: user text (from a jira issue) mapped to the team responsible for the fix. Can you share some tutorials? I only just managed to get this working on windows/cuda: https:/&
456.
▲
by
eurekin
3y ago
Yeah, same experience here
457.
▲
by
eurekin
3y ago
Ok, impressive! What are real world use cases for 7B family of models? Is anyone using them for anything productive?
458.
▲
by
eurekin
3y ago
Just chiming in to say thank you for doing such a product. Have you ever consider reaching out to any mirrorless manufacturer (maybe some form of a partnership?) about recording it's camera's sensor data? I have a Nikon and I'
459.
▲
by
eurekin
3y ago
I really don't know, if that even is possible. That's main reason, why cinema cameras are picked. I suspect that it could be possible now, to an extent. We have quite good image restoration tools, some based on neural networks. Ma
460.
▲
by
eurekin
3y ago
Probably auto-exposure adjusted the shutter speed to compensate
461.
▲
by
eurekin
3y ago
No attempt at real answer, but some hints from watching youtube videos on the topic: Lightning: - Edge or back lightning, if dramatic - Wraparaound (cradle) lighting, if for pleasantness - Low key look for interiors (no white walls) - Artif
462.
▲
by
eurekin
3y ago
I wouldn't be surprised if system token included "this photo contains following text: <output of tesseract ocr here>"
463.
▲
by
eurekin
3y ago
It advertises, it's for php development. Is it flexible enough to handle any http app? Does is support websockets, or maybe arbitrary tcp connections (such as jdbc postgres or activemq broker)?
464.
▲
by
eurekin
3y ago
proxmox/virtualbox/qemu + throwaway vm
465.
▲
by
eurekin
3y ago
Oh my.. Please, can you share more about it? Some NDAs about the function of that pi fleet?
466.
▲
by
eurekin
3y ago
Same. Went with Odroid HS3+ instead and couldn't be happier
467.
▲
by
eurekin
3y ago
One of aspects is putting latest knowledge to the model. They could, given enough requests, provide a special subscription service, where you get the latest model every N DURATION time. Or, maybe a finetuned version for your particular data
468.
▲
by
eurekin
3y ago
Interesting. I assumed that if rocks can melt, so would any substance given enough heat. Granted, not without chemical interactions (I assumed burning wouldn't count), but with protein they would at least unfold (denaturate). This pape
469.
▲
by
eurekin
3y ago
Theoretically, in a ozone-less environment and providing they don't sublime first
470.
▲
by
eurekin
3y ago
Was that the "how knowledgeable person that must've been" moment for you, before you noticed? :D
471.
▲
by
eurekin
3y ago
It's not uncommon to have it break partway through the response
472.
▲
by
eurekin
3y ago
Anecdata confirmation: I've been toying around with LLMs for simple fun stuff, but when it comes to real work, GPT-4 delivers in spades. I have cut many hours of debugging thanks to it. I could find issues easily, on-call in short conv
473.
▲
by
eurekin
3y ago
> 3) I'm leaning towards a Supermicro H12SSL-i motherboard which I've seen people run quad A5000's on without any hassle. Proper PCIe x16 slots spaces 2x apart, and i'd likely be watercooling all 4 cards together in a
474.
▲
by
eurekin
3y ago
Sure, can you create a bending (momentum?) diagram for that course? :D
475.
▲
by
eurekin
3y ago
To see the blobs forming, moving, changing
476.
▲
by
eurekin
3y ago
Would be cool to see those splats in action from a second, free to move camera
477.
▲
by
eurekin
3y ago
AFAIR, SVMs have one optimal solution, achievable analytically. NN can get stuck in local optima. If a transformer is a SVM, could we simply extract it out and optimise the hyperplane like for any SVM?
478.
▲
by
eurekin
3y ago
I have no answer, but if anything worked for you in the end, I would greatly appreciate to share what that was
479.
▲
by
eurekin
3y ago
I use one Windows desktop for a LLM & Stable Diffusion server (among others). I like to conserve power, so I have remote power switch (just a smart plug) and configured wake on Lan with very short time to sleep. I have been very happy
480.
▲
by
eurekin
3y ago
Are those kinds of algorithms used in VR eye tracking systems?
More ›