sevagh

20 ms·

301.
by sevagh 3y ago
Can you clarify your comment - you're referring to everybody who wants to work from home as "self aggrandizing adult children"?
302.
by sevagh 3y ago
I even feel the opposite would be terrible - imagine taking payment for something that started as a hobby project. Now the "pressure from users" is an actual contractual obligation! I can't walk away. What a nightmare.
303.
by sevagh 3y ago
>Are there any open source boosters left, and if so how would they make the case for open source at this point? I consider the way I publish MIT projects to be mostly in good faith. I put it out there because I like having my name associ
304.
by sevagh 3y ago
>free tech support Like, what, open-source police are going to arrest you for not addressing support questions? MIT/BSD is "free as in the maintainer is free to ignore any and all tech support requests."
305.
by sevagh 3y ago
Tyler the Creator famously tweeted "Hahahahahahahaha How The F* Is Cyber Bullying Real Hahahaha ** Just Walk Away From The Screen Like ** Close Your Eyes Haha." Seems like a ridiculous statement (and definitely doesn't addres
306.
by sevagh 3y ago
>I think the issue here is that we don't have a system where being an OSS contributor is a sustainable career path. They knew that before they started.
307.
by sevagh 3y ago
Reminds me of magenta/ddsp - vaporware.
308.
by sevagh 3y ago
If AMD 10 years ago tried to compete with NVIDIA and Intel both simultaneously, would it have survived? They did well to stay in business, and now they can address the AI thing.
309.
by sevagh 3y ago
Not that I maintain any hugely popular deep learning code, but my own "hobby"/dev hesitance at adding multiple devices is that I don't own them to test them. If I have an NVIDIA GPU, the only PyTorch backend I can promis
310.
by sevagh 3y ago
That's Canonical's problem.
311.
by sevagh 3y ago
>If you actually want to learn something about CUDA, implementing matrix multiplication is a great exercise. There is SAXPY (matrix math A*X+Y), purportedly ([1]) the hello world of parallel math code. >SAXPY stands for “Single-Precis
312.
by sevagh 3y ago
In this case the one singular event of being kicked out at a specifically vulnerable age of 17-18. The rest of their life was influenced by that.
313.
by sevagh 3y ago
In this thread I feel like I'm among my people. This year is the one where I've been reflecting. Over the past N years of being focused on self-improvement and perfectionism and a self-evaluation/grind/progress loop in q
314.
by sevagh 3y ago
Much less, much much less. Imagine paying $4,999 USD for a device from a company with $2.8T market cap and blaming anybody except Apple for the drawbacks of your platform.
315.
by sevagh 3y ago
Mac users are voluntarily in the walled garden of a _trillion dollar company_, it's a bit different from Linux.
316.
by sevagh 3y ago
I can't think of a more black and white situation than this. If the OP, Ariel, did not do what they did, the kernel would not be fixed. Therefore, they contributed the fix.
317.
by sevagh 3y ago
Sorry about that - could you put the contents of the developer console or the symptoms in a github issue (or describe it here?)
318.
by sevagh 3y ago
Can you create a github issue? Dod anything show in the developer console? I could (probably) improve umx.cpp but freemusicdemixer.com is supposed to be the "effortless" compile-free version of it.
319.
by sevagh 3y ago
Nice idea! The name 'other' comes from the traditional datasets used to train these models, but I'm open to trying to incorporate a better name.
320.
by sevagh 3y ago
Thanks for the kind words!
321.
by sevagh 3y ago
I envisioned my site would be useful for quick and dirty prototyping - especially since the outputs are not licensed for commercial use. Then, when one's project idea is validated, they could then used a paid stem separation service wi
322.
by sevagh 3y ago
Demucs is awesome no doubt. >plus you can run it with GPU support as well! Open-Unmix also originally runs on the GPU like it was intended for, since it uses PyTorch just like Demucs. I'm curious about using WASM + WebGPU to add GPU
323.
by sevagh 3y ago
Oh no! Can you open a github issue with the size of track you tested? Or email the track to me at sevagh@protonmail.com I can make the segment size smaller (right now I'm using 60s segments).
324.
by sevagh 3y ago
Hmm - that should be the end of it, at which point download links will appear to either the Single or Batch apps, depending on which one you used: - bass, drums, vocals, other, karaoke.wav in the Single track window (at the bottom: Demixing
325.
by sevagh 3y ago
Maybe I can ask HN out loud how I can start figuring out how to get this website on Google search results - e.g. I want it to show up when I search relevant terms like "free stem separation" "free music separation" &quot
326.
by sevagh 3y ago
That's awesome! I first got in contact with Fabien-Robert ( https://github.com/faroit ) during the Music Demixing Challenge 2021, where the UMX-L weights were first unveiled. We have since discussed my projects a couple
327.
by sevagh 3y ago
My dream (and next major project) is to get Demucs (v4, Hybrid Transformer) running in WASM, similar to what I've done with Eigen/C++ and Open-Unmix.
328.
by sevagh 3y ago
I'll be honest with you - I wouldn't even know what to do with user data if I collected it. I'm not great at training neural networks or understanding how to augment training with additional data. Basically, the _only_ specia
329.
by sevagh 3y ago
In my first post, quite a lot of alternatives were discussed: https://news.ycombinator.com/item?id=36707877 The model I'm using is called Open-Unmix ( https://github.com/sigsep/open-unmix-pytorch ).
330.
Show HN: Improved freemusicdemixer – AI music demixing in the browser(freemusicdemixer.com)
160 points by sevagh 3y ago|40 comments

More ›