bonoboTP

18 ms·

451.
by bonoboTP 4mo ago
These are vibes. ViT has been shown to work fine on small data with proper hyperparam and most of what you mention is actually doable just fine with the other architecture as well.
452.
by bonoboTP 4mo ago
Exactly. Most of the comparison papers are useless. This is hard stuff, only few people have the chops it takes to even attempt this. You can of course train some models and then post the numbers, that's not the hard part.
453.
by bonoboTP 4mo ago
What's the use case enabled vs running a ConvNeXt or EfficientNetV2 and using the resulting strided features as you would the resulting tokens of a ViT? I'm not saying that ViT is worse. Just saying that the scholarship around com
454.
by bonoboTP 4mo ago
Yes but not based on rigorous comparison. I'm not saying ViT is bad. But it took over mainly because it's the shiny new thing. It very bandwagon-Y even among PhD students.
455.
by bonoboTP 4mo ago
There's a dearth of research properly comparing them.
456.
by bonoboTP 4mo ago
You can run a CNN and use the downsampled feature map the same way as patch tokens.
457.
by bonoboTP 4mo ago
CNNs are fine when trained with a good recipe. There are very few good studies comparing them with proper hyperparam search and all the training tricks applied consistently. Transformers are good but ViT vs CNN is not some settled issue. Tr
458.
by bonoboTP 4mo ago
Why are you not on thunderbird yet? Why do you get Windows notifications? Are you using Windows? I don't understand how there are people who can notice such things but still use windows in 2026. Also, please don't write with AI. T
459.
by bonoboTP 4mo ago
Exactly. All these ideals work in theory but then in reality banks are also incompetent and will use all kinds of domains. Same with meta and Google where they often direct you to domains that aren't under their main one and it's
460.
by bonoboTP 4mo ago
Why can't it do that? How do you do it then? An LLM-based agent can execute any commands in a terminal, including running whatever commands that will send commands to an external device through whatever wires you have set up. Is the cl
461.
by bonoboTP 4mo ago
Ok but it takes weeks of trial and error for expert humans. If you allow the llm to run the code and see the results and do trial and error, it will also likely figure it out, no?
462.
by bonoboTP 4mo ago
Yes. We will likely have a heavily policed narrow line to some kind of AI during the phase when we as employees will still have value in this centaur configuration, so we'll be allowed to steer the largely autonomous models for our emp
463.
by bonoboTP 4mo ago
Did you try this by giving it access to the materials? Human programmers also don't memorize all this stuff. If this is the reason for your calmness it's quite shortsighted. There are problems when you rely too much on AI generate
464.
by bonoboTP 5mo ago
Meta made Pytorch and a lot of vision models back in the day, like Faster RCNN, Mask RCNN, the Detectron framework, and more recently the SAM and DINO series. AI not just LLMs.
465.
by bonoboTP 5mo ago
Yes, you need a happy medium. I know complainers who do zero socialization, never stay for social events even once per month, never join a Friday beer (sure, you can drink alcohol free) etc then complain they are overlooked. Humans are huma
466.
by bonoboTP 5mo ago
By doing it. Decide on a small project, like tracking your cat, detecting food items in your fridge, then take it step by step. Then do a slightly more ambitious project. Start with something very simple. It also heavily depends on what you
467.
by bonoboTP 5mo ago
Yes, but it wasn't invented from nothing in 2017. Soft attention existed in other applications like information retrieval, Nonlocal networks had similar ideas as well. But it wasn't seen or used as a fundamental building block. Bu
468.
by bonoboTP 5mo ago
I didn't realize that was what you're thinking about. Well yeah, those guys invest in social connections at work, while you are not. Alliances, trust and human relationships will always matter, no matter how much people complain a
469.
by bonoboTP 5mo ago
That's the right question. I don't like this dichotomy between domain and engineering. It seems to come from people who just build different CRUD apps for mobile and websites for businesses in different industries and that's
470.
by bonoboTP 5mo ago
True of human-made things as well. Most video essays don't get more than a dozen views, most gameplay streams similarly. People playing their guitar and uploading, same. SoundCloud, YouTube, twitch. Human-made app store apps is the sam
471.
by bonoboTP 5mo ago
There will always be people who are willing to put in the time and grind a few years in their 20s to achieve a better life for themselves and their future family, even if you would prioritize other things. It's ge theory and you don&
472.
by bonoboTP 5mo ago
Digital files that you store on your own storage media with free software also can't change (without your intervention). But in new generations many only have phones, not even laptops.
473.
by bonoboTP 5mo ago
Attention layers were not used in the 90s.
474.
by bonoboTP 5mo ago
It's hard even at the highest levels, such as in writing scientific papers or doing scientific conference talks. People just generally have a hard time to step outside of their context and think with the head of someone who has a diffe
475.
by bonoboTP 5mo ago
Well, today's multimodal llm agents with tools would at least have a good chance to do something with even such an underspecified query. Because fixing things is simpler to specify, the agent could look at config, network settings, sen
476.
by bonoboTP 5mo ago
Going from one off prototype to robust product is a huge leap. I think these ephemeral context tailored projects are really great and useful. But these are not to be thought of as products. They work for you specifically, and people who are
477.
by bonoboTP 5mo ago
I have some friends who, since their high school days help some older acquaintances in upgrading their PCs, choosing laptops and phones, helping with setup etc and these older folks have comfortable money and pay him very well above what wo
478.
by bonoboTP 5mo ago
Opus 4.5 fixed so many issues with my self-coded research projects, and allowed me to port between tensorflow and Pytorch in a much shorter time than manually. Helped a lot with docs too.
479.
by bonoboTP 5mo ago
Some people are really bad at specifying what they want to ask for. Or they already start prompting with the attitude that it can't possibly work so they don't even really try, or stop at the first failure to point and say how bad
480.
by bonoboTP 5mo ago
I think the GP just means it was their oh shit moment, not that it was their image...

More ›