A colleague of mine has written up a quick explainer on the key features (https://encord.com/blog/segment-anything-model-2-sam-2/). The memory attention module for keeping track of objects throughout a video is very clever - one of the trickiest problems to solve, alongside occlusion. We've spent so much time trying to fix these issues in our CV projects, now it looks like Meta has done the work for us :-)
Those are both great tools. However, there are a number of differences, but the two most prominent are that: 1) Encord Active automatically analyses internal metrics to find the most relevant data and labels to focus on to improve model performance; and 2) it is optimised for the full 'continuous' training data workflow including the human-in-the-loop model validation and annotation.
We base our pricing on your user and consumption scale and would be happy to discuss this with you directly. Please feel free to explore the OS version of Active at https://github.com/encord-team/encord-active. Note that some features, such as natural language search using GPU accelerated APIs, are not included in the OS version.
We run usage-based and tiered pricing, but we haven't gotten around to building out a self-serve "sign-up-with-credit-card" product yet. For all the advances in Stripe and automated billing, these things still take some time to implement for a short-staffed engineering team :-)
Yes - the tool can definitely help with that. We combine the newest embedding models with various other heuristics to help identify performance outliers in your unseen data.
It takes me 2 seconds to switch to a different design tool. Adobe has 30K employees and been around for 30 years. Figma is a small scrappy startup. It seems pretty preposterous to me that the deal will have any meaningful impact on consumers.
You don't have any knowledge, skills, habits, techniques or workflows specific to the tool you use? If true then your use is... anomalous and you shouldn't present yourself as a typical user of this tool.
They also have an insane amount of traffic - very envious, they'll be able to capitalise on other things as soon as they start to build commercial products people are willing to pay big money for.
How do you plan to expand the functionality? I've been looking for a more modern alternative to OpenCV for a while - especially replacements some of the aged object tracking functions, there's been a wad of new stuff coming out that OpenCV has been slow to implement.
Great question! Development has been guided by my asking myself -- and discussing with friends -- what problems I/we would like to solve (or what would be fun!), then building for those use cases. This is evident in the WIP DetectDistinctScenes[] method.
This method is more advanced than most. It uses CLIP behind the scenes to detect when classifications change for more than N frames in a video, and generates a list of timestamps you can use to determine scene changes.
I'm keen to build more specific functions for common CV tasks. My code most certainly needs cleaned (the main lang.py is > 2k LOCs right now) but I do have a contributing guide that you can follow to add features should you be interested: https://github.com/capjamesg/visionscript/blob/main/CONTRIBU...
Shameless plug: take a look to our embedded computer vision library SOD: https://sod.pixlab.io.
It's a lightweight OpenCV alternative targeting embedded devices with most of the modern image processing algorithms already implemented including an experimental Stable Diffusion implementation.
a new, popular library for basic functionality (converting between annotation formats, evaluating models, doing object tracking) has been supervision: https://github.com/roboflow/supervision
Their traffic is rumoured to be down 25-30% after ChatGPT was released, so good to see that they are doing something to keep the community alive in the new era of GenAI.
They're currently sending, massive unsolicited e-mail campaign to all (sic!) stackoverflow users. Message they're sending below. Please find 6 places where they try to address AI directly and indirectly :)
Stack Overflow is investing heavily in enhancing the developer experience across our products, using AI and other technology, to get people to solutions faster.
As part of that initiative, we’ve launched Stack Overflow Labs.
Here is where we’ll share our experiments, demos, insights, and news - across all Stack Overflow products. We plan to continually add to this site as we experiment and release new solutions.
You can sign up to get previews and early access to features that will be available on stackoverflow.com.
Our guiding principles for Stack Overflow Labs
Find new ways to give technologists more time to create amazing things.
Accuracy is fundamental. That comes from attributed, peer-reviewed sources that provide transparency.
The coding field should be accessible to all, including beginners to advanced users.
Humans should always be included in the application of any new technology.
Very interesting -- had not seen that. Indeed, I think they could do with a bit of a makeover. Their ad platform is also terrible, way worse than Twitter's, so it is both good and bad that they're feeling a bit of a pinch I suppose.
more than that the platform is the more aggressive in the worst way possible, or your ask so specific about weird behavior that nobody responds or they say this already bean asking even if not(because some mods don't read what they ban), they block conversation with hundreds of votes, they practically insult the post writer for “not knowing”, i'm happy if they start changing the nefarious moderation scheme, im 100 % more likely of making the same question in reddit than in stackoverflow people are kind, because moderation encourage that, that would totally increase the use of their platform lots more than this.
how do i remove diacritics from unicode characters https://www.phind.com/search?cache=bd2b33eb-9454-4d38-975e-1... and it answered with multiple Python code snippets with the second solution getting close to the real solution but is slow and incorrect.
Now if I ask the question I actually want
how do i remove diacritics from unicode characters with php https://www.phind.com/search?cache=c29fe466-16cc-4795-94bf-0... then you can see it misunderstanding the question: the first answer is the desirable output (I suspect it only works for latin letters tho) except for the iconv problems mentioned but the second is completely incorrect. The third answer is close but no cigar as it only works with Latin characters.
I'm not getting great results. It seems to love to latch on to a keyword in the question and run with it (maybe the GPT-4 version does better). I asked a question about async in Rust, it gave me a comparison between threads and async and then kept going in that direction instead of answering what I asked. I mentioned a specific crate as something I wanted an alternative to, and it told me a description of the crate. Rephrasing the question a few times eventually got a correct answer, but it's not much better than going straight to ChatGPT.
I've tried Phind a few times, but my gripe with it, so far, is it gives me outdated examples that don't work because the API has changed. Is there some magic keywords I should be including to prevent that?
It's far from perfect, but in my experience, it's much better than both ChatGPT and Bard. I've also used it to ask non-technical questions, and it often generates a good response, so not really just "for developers".
25% of that traffic was asking questions answered by a glorified markov chain. I dont think stackoverflow lost any valuable users or interaction unless they're looking for newbies with syntax errors to flood the place.
I think they are looking for that. As long as they read the ads, they don't care how good at programming they are. It's a "web property", not a programming contest ;)
