ex3ndr

7 ms·

31.
by ex3ndr 3y ago
This is a very weird thing to say when expo is literally one line installation and it does not require you anything to be installed (even XCode) on machine for most projects. It is literally 10x times easier than just writing in Swift.
32.
by ex3ndr 3y ago
Because they run JS on main thread while react native tries to minimise it at all costs (only keeping rendering and animations, but not layouts for example).
33.
by ex3ndr 3y ago
IR sensors are just cameras, also IR can look through some clothes
34.
by ex3ndr 3y ago
No, they have they are just "IR" cameras and they avoid using word "camera", but they are still pretty much the same cameras.
35.
by ex3ndr 3y ago
Current roborocks are mostly vision plus some lidar for some precision
36.
by ex3ndr 3y ago
And yet native apps consume as much or even more memory than electron because you have to implement memory management manually which could be very difficult for app like WhatsApp.
37.
by ex3ndr 3y ago
they released one quite minor modification to largest whisper model and in fact it much worse than a previous.
38.
by ex3ndr 3y ago
IIRC first version was very overengineered and team was replaced that built a simpler one.
39.
by ex3ndr 3y ago
you can buy pcie extensions easily on amazon.
40.
by ex3ndr 3y ago
4k is trivial, 5k is impossible, your link is about 4k.
41.
by ex3ndr 3y ago
I highly doubt that building a 10 c++ files was as fast as now. I am clearly remember building c# apps for a much longer, i remember how deployment of a trivial app on Azure took ~15 minutes. Meanwhile we still don't have 5k displays w
42.
by ex3ndr 3y ago
also they include llama.cpp, i guess they just download it instead of bundling with binary like ollama.
43.
by ex3ndr 3y ago
Just FIY, you have whisper.cpp in pipeline, but be advised that it is for some reason barely works. tiny model using transformers.js running via wasm outperforms any model of whisper.cpp. fast-whisper is much better here and as fast.
44.
by ex3ndr 3y ago
This is basically this - it can learn ignore some paths, and amplify something more important, then you can just cut this paths without sensible loss of quality. The problem is that you are not going to win anything from this - non-matrix m
45.
by ex3ndr 3y ago
Same here! One of my very first apps for android 1.1 was a port of bochs =)
46.
by ex3ndr 3y ago
it is barely relevant to big players, but is extremely valuable for small players: distributing your workflow to multiple gpus manually is not that simple thing to do and there are a lot of much more interesting/important problem to so
47.
by ex3ndr 3y ago
Wait, they explicitly allowed to use cards for blockchain and prohibited for anything else?
48.
by ex3ndr 3y ago
No, this cards would be inferior to 4090, that you can get everywhere and you can buy 3 of them for "spare parts"
49.
by ex3ndr 3y ago
This would end up memory-bound for LLMs (and you would better get x3 4090 for this) or by compute for diffusion models (and again 4090 is like 10% slower than H100 for 1% of the price).
50.
by ex3ndr 3y ago
it is 20-30% slower than 4090 or H100 in compute, the only improvement is slightly more RAM. This card is not for ML (on purpose) - it is for more enterprisish tasks: some advanced video streaming/rendering, virtualization, etc.
51.
by ex3ndr 3y ago
Very disappointing, much less cuda cores than in 4090, no nvlink (32G won't be enough anyway), price x3 to 4090.
52.
by ex3ndr 3y ago
They literally mention that they built on top of llava that was released half year ago.
53.
by ex3ndr 3y ago
Do you really need a vibe to build though?
54.
by ex3ndr 3y ago
Like what? Most countries in europe still don't have transfers by phone number?
55.
by ex3ndr 3y ago
I was picturing more like VW ID buzz
56.
by ex3ndr 3y ago
So a bigger and with much higher bonnet than SUV? what's the point then?
57.
by ex3ndr 3y ago
Agree, it is bizzare how all this discussions completely ignore the children and families that have to get decent sized car.
58.
by ex3ndr 3y ago
I just tried a lot and the best thing you can get is to do something practical (most ML is empirical anyway) and pick something that you can train on small machine. I picked working with audio since it usually don't need too much data,
59.
by ex3ndr 3y ago
So running on M2 Ultra would beat 4090 by 30%? (since it has 2x of gpu cores)
60.
by ex3ndr 3y ago
Very useful thanks! I guess i would just buy them then.

More ›