naillo

30 ms·

1. 331.▲
   Optimizing models using the PyTorch JIT(lernapparat.de)

   2 points by naillo 4y ago|0 comments
2. 332.▲
   by naillo 4y ago

   Not creepy at all
3. 333.▲
   by naillo 4y ago

   Maybe they should be doing continuous UI latency testing
4. 334.▲
   by naillo 4y ago

   Huh? As long as they have pictures, they can just run a classifier on it to figure out if you have couch etc. SLAM or not doesn't matter.
5. 335.▲
   by naillo 4y ago

   I had such a different experience. I found the EULA surprisingly readible and easy to understand. Just 5 easy bullet points rather than a multi page document of legalese.
6. 336.▲
   by naillo 4y ago

   I'm not sure this is the standard. I have a feeling Keenan Crane is trying to set a (fantastic) example for others to follow. His presentations are really top notch and something everyone should strive for, a really inspiring smart guy
7. 337.▲
   by naillo 4y ago

   Idk. I just downloaded this for kicks and I find VLC a lot more snappy. I definitely felt noticable lag when switching between tabs with this.
8. 338.▲
   by naillo 4y ago

   I get that people do it, it's just that I've never gotten the advice to do it (non body building contexts, just normal life). To "småäta" in sweden is a widely talked about thing to avoid (i.e. adviced against).
9. 339.▲
   by naillo 4y ago

   > following the advice of eating small meals often This must be an american thing. I've always been told the opposite advice all my life. Interesting.
10. 340.▲
    by naillo 4y ago

    If you switch to julia you'd have to port over the entire pytorch ecosystem as well if you want to use this together with that (taichi has great compatability via `to_tensor`). Also, the kernels you write in taichi is hardly a DSL, the
11. 341.▲
    by naillo 4y ago

    Old enough to remember when the accelerometer didn't require permission to access.
12. 342.▲
    by naillo 4y ago

    Makes me think of Carmacks remark (from the lex fridman interview) that he 'almost count anyone out' that tries to make AGI happen through real life humanoid-ish machines. (Meaning simulated worlds that can do e.g. 1000x real time
13. 343.▲
    by naillo 4y ago

    Is this really using stable diffusion? Has a really 'midjourney' type style to it.
14. 344.▲
    by naillo 4y ago

    Which is so silly since ML models should be the most portable thing in the world. It's just a series of math operations, not a bunch of OS/hardware specific API calls or something like that. We should be at a stage where each ML m
15. 345.▲
    by naillo 4y ago

    Wonder what people who like to employ the "it's their website, they can do whatever they want" argument feels about this.
16. 346.▲
    by naillo 4y ago

    Isn't this what infinite jest was entirely about
17. 347.▲
    by naillo 4y ago

    I thought we'd never get image generation this fast. Last year it was 30 minutes per image. The stable diffusion folks are planning for a 100mb release of the image generator in Q1 which for sure would be real time. I actually suspect
18. 348.▲
    by naillo 4y ago

    Imagine how fun sitting at a terminal in vim editing a 100 line 'script' for a short movie and getting rapid feedback back. I'm so excited about the future.
19. 349.▲
    by naillo 4y ago

    I kinda feel like they chose the name "open"ai when they started back in 2015 because musk etc wanted exactly the kind of thing stability ai is now creating. I.e. something other than a corporation like google having primary acc
20. 350.▲
    by naillo 4y ago

    I can't prove it of course but it wouldn't surprise me if they had this pretty much done already long ago (dall-e has been out for several months at this point). The actual implementation doesn't look like it'd take more
21. 351.▲
    by naillo 4y ago

    I can't help but feel like they're adding this at this particular point since Stable Diffusion has announced they're releasing their 'inpainting' model next week.
22. 352.▲
    by naillo 4y ago

    Try it yourself ;) I haven't spotted any but I can't dispute the parents observations.
23. 353.▲
    by naillo 4y ago

    Just throw it into im2im stable diffusion (joking but it also would likely literally work).
24. 354.▲
    by naillo 4y ago

    Might be that there's some degredation but I think it's pretty close. Anyway I'm using their 'official' fp16 version which they might be doing some extra magic on idk. I.e. via StableDiffusionPipeline.from_pretr
25. 355.▲
    by naillo 4y ago

    Exchanging compute for space in the space-compute tradeoff curve.
26. 356.▲
    by naillo 4y ago

    I found this one to be really rad too: https://www.reddit.com/r/StableDiffusion/comments/x0u9a8/ama...
27. 357.▲
    by naillo 4y ago

    Not really relevant bc it's kinda nitpicky but amazingly it's not even 4gb, it's more like 2gb if you use the float16 version (which has no quality degredations). Quite amazing that so many images fit in that small a package.
28. 358.▲
    by naillo 4y ago

    The parent is about im2text not text2im (i.e. converting an image to human readible text). Easy thing to misread. Apologize for upsetting you. Have a good day :)
29. 359.▲
    by naillo 4y ago

    Are you literally a bot? This is the second time I've seen you link to replicate.com (this time the link isn't relevant to the parent).
30. 360.▲
    by naillo 4y ago

    There's like 5% of people who actually use adblock too. Seems rude to go out of their way to just anger people like you.