@ivoflipse

11 Followers
66 Following
31 Posts
@savannah I think with some.of the discussions going on right now, they're in desperate need in someone rallying everyone around the right things to work on and you seem like a great fit! Now just to find the funding somewhere...
@savannah naah, you need to get in over your head even deeper. They should hire you to be the PM for CPython 😉

#Python and #PyTorch using #WebGPU in browsers, the future is now! 🤯

https://praeclarum.org/2023/05/19/webgpu-torch.html

How I Re-implemented PyTorch for WebGPU

TL;DR I’ve been working on a WebGPU optimized inference and autograd library called webgpu-torch with an API that matches PyTorch. The goal is to run neural networks in the browser at speeds comparable to a Linux workstation. Many kernels have been implemented and its design is easily extensible. It’s available on NPM now and works in both the browser and Node.js!

praeclarum
@willmcgugan could be a very nice widget for viewing dataframes with pandas or polars
I have found the best ever TikTok video and it's honestly not even close https://www.tiktok.com/@unknowndazza/video/7204587065726520577
Dazza on TikTok

The Full Frod Story! It’s been a crazy journey from 0 to 60k people of Frodrick! Appreciate your ideas and support☺️ #Frodrick #startedfromthebottomnowwehere #FrodHouse

TikTok

I’ve been working on pygfx, a Python graphics library based on wgpu (the successor to OpenGL). I wrote a post on it's background https://almarklein.org/wgpu.html

This project is close to my heart, visualisation is what got me into Python. For scientific Pythonista’s pygfx might still a bit low-level - the idea is that others can build high-level libraries on top.

#python #viz

On WebGPU, wgpu-py, and pygfx

@tonybaloney @hugovk based on these graphs, introducing a nogil change doesn't have to be too bad. It would take many years before mainstream Python users switch to Python 3.12, hopefully giving us time to deal with the incompatibility
@brettcannon @gpshead @jacob perhaps it would be good to mention this on Discuss? I feel like your point of view doesn't seem well represented there right now. This seems like the first PEP (that I know of) that would be conditional on a certain level of funding, so it would be good to hear everyone's perspective. Perhaps the same applies to the packaging discussions, which efforts could benefit from funding
@brettcannon @gpshead @jacob we as a community, since you also maintainers on board to make or accept the changes, folks to test packages to find any problems, core devs to deal with the consequences for the stdlib. When you say fund this work, are we talking a dedicated 4Developer in Residence or what kind of price tag do you have in mind?
@brettcannon @gpshead @jacob As with lazy loading, the uncertainty of existing code supporting or being compatible with these new features is a serious concern. But if we don't offer a way to try out such features, how are we going to get past this obstacle? What incentive would projects have to do something about it?