AFAIK, only CoreML can use Apple's NPU (ANE). Pytorch, MLX and the other kids on the block use MPS (the GPU). I think the limitations you mentioned relate to that (but I might be missing something)
I totally agree with the post.
The definition of a museum is "an institution dedicated to displaying or preserving culturally or scientifically significant objects", according to Wikipedia. Most of the time I do not see anything significant on these screens in museums, since equivalent content can be easily reached on any phone.
Real, relevant objects are much harder to find and find a way to create interest around them. But that is exactly what makes a museum a good museum, not the screen.
Very nice! I wish they were more keen to contribute to AI/ML community an publish weights and model definition on HuggingFace.
Funny enough I have just seen today a similar demo that is using a freely available VLM: https://github.com/ngxson/smolvlm-realtime-webcam
It seems nice. When programming envs tend too complicated though, I tend to switch to docker: a given image will handle a given version and it's mostly hassle free. That probably would not make sense on macOS or Windows, if you want to run native versions.
This is a common thing in software too I fear. I have Firefox on macOS. Everything is configured for US English, but I live in France. When I tell Firefox to save a page (or download something), the dialog prompts me to "Enregistrer sous..." (i.e.: "Save As..."). No idea why this happens.
It should be possible to see if a given "secure" binary makes use of a given microcode instruction. If that is the case, it could be possible to use udbgrd/udbgwr to modify the microcode instruction to do something else, e.g. to store some register data in the URAM or the SRAM. By doing this you should be able to "spy" the data the "secure" binary is processing.
https://brainchipinc.com We are looking for a C++ software engineer with 5+ years of experience to work with a small team to design and build software products that are part of the Akida Development Environment.
The Akida Development Environment is an easy to use, complete machine learning framework for the creation, training, and testing of neural networks, supporting the development of systems for Edge AI on Brainchip’s Akida event domain neural processor.
https://brainchipinc.com - Brainchip is searching a C++/Python software engineer with 3+ years of experience to work with other team members to design and build software products that are part of the Akida Development Environment.
The Akida Development Environment is an easy to use, complete machine learning framework for the creation, training, and testing of neural networks, supporting the development of systems for Edge AI on Brainchip’s Akida event domain neural processor.
https://brainchipinc.com - Brainchip is searching a C++/Python software engineer with 3+ years of experience to work with other team members to design and build software products that are part of the Akida Development Environment.
The Akida Development Environment is an easy to use, complete machine learning framework for the creation, training, and testing of neural networks, supporting the development of systems for Edge AI on Brainchip’s Akida event domain neural processor.
One thing I wanted to do for a project at work is to add python bindings to our c++ library, having the least effort for having a decent python binding. Two features I am looking for are support for return values of custom objects of arbitrary size and c++ exceptions raising python exceptions. Cppyy seems to support these, do you know if/how they work on pybind11?
I had always seen HN as a place where people curious about technology, culture and society share article links and comment about them. The moderation is what makes it work. I really do not see why censorship (or detox if you prefer) on political matters should be necessary.
if HN really needs this, it might even mean that its audience needs to be educated. And if that is the case it might mean the level you get in here is deemed to decline.
The top change I would love to see in C is to have a way to have conditional macros, that could produce efficient statically compiled chunk of texts of defines. I think you can get some of that in C++ with templates, but without the whole C++.
The parent comment is really a good advice. I really think that doing an activity in a group helps: when you get too shy you can comfortably hide in the group, when you want to talk you will have people with similar interests that will be keen to listen.
Now I am in my 30s, but when I was 24 I left my home country in Europe to work for an year and a half in the US, then moved again to another foreign country (UK).
Work was great, but it was really difficult to meet new people, not to mention girls. My English wasn't good and I didn't know anybody apart from my colleagues. So I learned it the hard way: I enjoy photography and I got into a local flickr group. Doing photo walks allowed me to get in touch with different people with the same passion as me, without having to think too much as the way I used to speak. The "walks" allowed me also to walk by with one person at the time, avoiding the awkward moments when you are supposed to talk in a table where all the other people can talk easier than you. I also met a good bro who I still see quite often! And when bored, I could ask him if he wanted to go to the cinema, to have a beer or to see an art exhibition. Few years later we even got into talking to some girls in bars. That didn't go too far most of the time, but it helped the self esteem and I could feel better when talking to random people.
I also started rowing (but you could try another team sport). It was a beginners' team, and even if I left it in the end I really think it boosted my confidence to hear the other guys support.
Also, I really think you can improve your English by watching TV and films. And learning to drive/having a car might inspire you to join interesting activities (my experience says that the US can be a difficult country to live in without a car).
I worked on the development of the P7, and yes, it is a dual core CPU powered by ARM cores with a quad-core GPU. The CPU is meant to be used for other projects too, so a quad core GPU makes sense.