HN user
A PyTorch-based musical mel-frequency transform for audio processing, optimized for performance and ONNX compatibility.
If you've ever wanted audio features that directly represent semitones (or quarter tones!) this is the package for you.
very cool. how are you creating the control vectors? curious since the word "cold" can be both a conversational disposition and a temperature (same word)
Of course not! :)
The Ray ecosystem is actually chalk full of integrations, from XGBoost Ray (https://docs.ray.io/en/master/xgboost-ray.html), to PyTorch on Ray (https://docs.ray.io/en/master/using-ray-with-pytorch.html), and of course hyperparameter search with Ray Tune for a variety of libraries, including Sklearn (https://github.com/ray-project/tune-sklearn).
For Python I'd recommend: https://github.com/worldveil/deuces
I have Comcast. I did it with this guide: https://willdrevo.com/blocking-ads-with-pihole-mac-os-x-ipho...
I would love to see an ELI5 writeup of this. For instance, I can't quite warp my head around - how are they storing neuron state in the lattice?
what kind of DSP were you doing? can I come work for you haha
Tile images, set randomness, build videos with tile images, and do face alignment montages with embedding network:
this is super cool. do you think the primary demand driver for your sales will be 1) gov't mandates, or 2) cattle farmers wanting to differentiate their meat to eco-conscious customers / retailers? just curious about how you see the market.
Right. The opacity setting is superimposed after the tile assignment. Thus you can have both.
I’ve found that opacity if 0.7 is often a nice compromise.
Both! You can experiment with —-best-k and —randomness <1.0 and sort of get things in the middle.
Basically what this will do is assign each tile less than 100% of the time randomly and then for each that isn’t, you choose among the top best K matches on L2 distance with equal probability. Gives it a little bit of both.
Totally. Everything improves with more images since that’s your palette for painting, so to speak. The ones I posted were with less than 100 images so you can definitely do better.
L*b was very slow, yes. It’s all done offline but I tend to like quicker feedback...
Right. See my comment above.
As to why it was in the project: If you treat each pixel as an example vector in 3 dimensions and cluster, you get the “dominant” colors for the image. It’s a primitive way to compress images as well. In this case I just was using it to generat fun cards that would use a minimal number of dominant colors. It’s still in the code if you’d like to use it but a bit hidden.
It didn’t make a ton of difference empirically when I tried it.
But you’re correct, generally that’s the space you want to be in.
Naw I had used the Kmeans part for extracting dominant colors. I was thinking about using it to generate cards and using the top K colors for the background of the card.
As I mentioned above the only ML is the face classifier. There’s a flag that allows you to only include face pics in the photomosaics as well
The face montage building trains a linear classifier on top of the pretrained embedding network - it’s the portion that talks about creating a training folder of your face.
But yes the photomosiacs strictly don’t use ML, unless you count the internal fun stuff Faiss (the similarity search lib) does to construct fast indexes.
This is super cool!
Do you have planned support for consensus-type mutlisig access? (ie: needing M of N approvals to acccess resources)
This guy needs to take a Stats 101 class. Taking a biased sample of Reddit posts and claiming that it must equate a liquidity crisis at the world's largest crypto exchange? Please.
80% of the problems Coinbase and every other exchange has are dealing with "Finance 1.0" banks & payment processors who still move funds around by having their employees manually type in numbers and sending transfer records via FTP. Seriously.
In a sense, Coinbase and other exchanges are moving money out of the traditional (fragmented) banking system and into a digital one. That takes friction and dealing with fragmented systems.
And personally I'd trust Coinbase over any of the other ones any day of the week. Have you dealt with Kraken/Poloniex/etc? Terrifying.
If you're using a free service with no ads, then of course they are selling any data you give them access to.
This isn't really news.
1) Does it integrate with my calls from my iOS / emails from gmail? So that I can sort by "least recently connected"? otherwise writing notes each time and looking it up gets tiresome. 2) is there a mobile app? 3) how much will it cost monthly? 4) how are reminder sent?
Just because you shuffle the examples on a single phone/user doesn't make it stochastic.
The entire point of using stochasticity (ie: random shuffling) is to avoid similar and/or a same-ordered run of examples from redirecting the hill climbing in a globally non-optimal direction all at once.
A single user's examples will be very similar, so you can shuffle all the examples from one user you want - that doesn't make it truly stochastic in the context of gradient descent optimization.
The quantization / compression part is pretty cool though. I suppose that could obfuscate slightly what the original example was for privacy purposes? Seems like you'd lose on accuracy though.
This is literally non-stochastic gradient descent where the batch update simply comes from a single node and a correlated set of examples. Nothing mind-blowing about it.
You realize that softmax scores aren't probabilities, right?
It's just a relative measure of confidence, scaled such that they all sum to 1.0.
Is this a library that does entity recognition and matching on individual elements in streams, or are the streams themselves being featurized/matched?
Does mypy see any performance improvements over vanilla python (because types are static)?
Exactly. Typically you won't see radio/streaming services messing around with reverb, but certainly different stereo systems have their own EQ (changes phase, amplitude) and often times the format/bitrate (again: phase, amplitude) or loudness (amplitude) will be different.
Most people don't realize this, but usually artists/labels will have a different mastered version of a song on each platform. Spotify, for instance, has it's own normalization algorithm that it puts tracks through to even out the listening experience in terms of loudness (RMS). Of course artists and their mastering engineers want to have some control over how that sounds, so they will change it.
Oh very cool! It looks like this runs in realtime as well?
Not really. It's not like a neural network which approximates a crazy function to the feature space leaving brittle points where you can exploit that to get false positives.
You could, for instance, insert other tracks into the audio additively to try and confuse the fingerprint retrieval logic into suspecting a different track, but since this and many other fingerprinting techniques depend on the actual frequency of the audio emitted, there's no shortcut to obscuring the actual track.
Oh, totally interesting. Didn't know that.
The American vernacular is you get "dejavu" whenever you have that "I've been here before..." feeling.