> Google is acting here as the owner and maintainer of a services ecosystem... > they increasingly experience difficulty to contain issues within that ecosystem and prevent them from spreading (piracy, malware, hacking,...)

I wonder if the smartphone app industry is big enough now that allowing just two corporates to govern them is no longer fair or democratic.

It has outgrown the "ecosystem" word a long time ago. It's a genuine industry now.

Apps are such a fundamental part of most people's lives now (whether they like it or not), and these two companies have a disproportionate amount of power over an entire industry.

This is more or less the journey the EU has started with the Digital Markets Act, but in a very agnostic way.

They identified that, among others, Apple and Google are operating a "Digital Market" of significant size within their ecosystem, where they invite others to participate and compete, and it's the role of a government to ensure fair conditions in the markets of its economy so forces can flow freely.

The way they defined that is very smart. They don't define what an "app" is or an ecosystem, they identify in a objective way that their operations constitute a market, and that they have to comply to certain rules in order to ensure fair competition.

I just hope that they can see this through and have those Digital Markets established as proper "markets" in the same sense as physical markets are, before some political "wind of change" is dissolving everything again.

Apple is very much counting on such a wind of change, by actively rallying its users against the EU...

https://pypi.org/p/torchruntime might help here, it's designed precisely for this purpose.

`pip install torchruntime`

`torchruntime install torch`

It figures out the correct torch to install on the user's PC, factoring in the OS (Win, Linux, Mac), the GPU vendor (NVIDIA, AMD, Intel) and the GPU model (especially for ROCm, whose configuration varies per generation and ROCm version).

And it tries to support quite a number of older GPUs as well, which are pinned to older versions of torch.

It's used by a few cross-platform torch-based consumer apps, running on quite a number of consumer installations.

I'm trying to understand this topic more, so I'm curious what you mean by the word tracking. I searched around, and wanted to check if my understanding is roughly correct:

Is tracking when the frequency data of the instrument, and the frequency+tempo changes in the music track are stored? And does midi just say "guitar" or "piano" and leave it up to the software to decide what those instruments sound like? So tracking would always reproduce the same sound, while midi can vary, even if it's making the same tones?

For e.g. an acoustic guitar and an electric guitar might both be producing a note at a particular base frequency (e.g. C2) but the overtones and amplitudes of those overtones would be completely different, giving each instrument their particular sound. So does tracking record that overtone distribution for each instrument, to ensure an acoustic guitar sounds like an acoustic guitar (that the music composer wanted)?

This is the source that I was reading - https://scalibq.wordpress.com/2017/03/29/trackers-vs-midi/

If not, I'd really appreciate any other reading material to understand what you mean by tracking, thanks!

People do use VR for modeling/animation in Blender - https://freebirdxr.com (disclaimer: I'm the author of this plugin)

Freebird is still quite basic, but it's under active development and more modeling and posing features are being added.

So a core set of people definitely use VR in Blender for creation/editing, but I agree the number is quite small.

I’m pretty sure the GP was thinking of something having qualities such as "funny", "well-written", or "creative".

This is like the least helpful or useful kind of comment ever. Hey I got barfed on by an LLM wanna see?

It runs locally if you download the "m" and "l" files alongside it (for maps and models), and run it on a localhost server (to fix the CORS errors).

curl "https://phoboslab.org/q1k3/index.html" > index.html

curl "https://phoboslab.org/q1k3/l" > l

curl "https://phoboslab.org/q1k3/m" > m

python -m http.server

--

I then zipped it up, and converted that zip file to 7 QR codes, so technically yes, we can fit Quake on a few QR codes.

So bad is not a full html self runnable, I guess it is doable with some quick trickery. In that case, one could build a QR code html interpreter (aka read > output to file > open the file with firefox ) and it would be so cool

I did that trickery and it works: https://github.com/AdrianVollmer/Zundler

(I had to fix the mime type detection and I'll have to push that fix to PyPI, so use the main branch if you want to try it yourself.)

Unfortunately, since we have to ship it with the gzip lib pako.js, the size more than quadruples to 77kb, so no QR code with this approach. Maybe I should look into making compression optional.

You can use DecompressionStream instead of pako.js for all modern browsers. FetchCrunch [1] uses it to generate self-extracting HTML files, which were not portable at that time but now fully portable across browsers.

[1] https://github.com/subzey/fetchcrunch

Very cool. Thanks for letting me know!

Uh uh thats very interesting. I am gonna play with it a lot.

> the size more than quadruples to 77kb, so no QR code with this approach.

OR we can use 25 QR codes haha. Joking, but what about huge QR codes?

This is cool.

I found that the first line didn't work but: curl "https://phoboslab.org/q1k3/" > index.html

did the trick...

Yeah, I assumed it was made by Comfy's author, as maybe a next evolution of their project. I was feeling happy for them.

But since it's not made by Comfy's author, it's definitely not cool to name your project by using their name.

Easy Diffusion (previously cmdr2 UI) can run SDXL in 768x768 in about 7 GB of VRAM. And SDXL 512x512 in about 5 GB of VRAM.

Regular SD can run in less than 2 GB of VRAM with Easy Diffusion.

1. Installation (no dependencies, python etc): https://github.com/easydiffusion/easydiffusion#installation

2. Enable beta to get access to SDXL: https://github.com/easydiffusion/easydiffusion/wiki/The-beta...

3. Use the "Low" VRAM Usage model in the Settings tab.

Hi, I'm the author of the cmdr2 UI and installer (that iFire linked to). I'd be happy to contribute the 1-click installer used by my project. It's a battle-hardened installer (over 100k installations on all kinds of PCs and networks), and I'm finishing up a rewrite in python, so that the installer code is easier to maintain for others.

I'd be happy to submit a PR to your project, if you're interested in using it. I actually got it working with your project a few weeks ago, so I know it works with your repo.

I've opened a github issue as well, so we can talk there if you'd like: https://github.com/invoke-ai/InvokeAI/issues/1042

Like I mentioned elsewhere, https://github.com/cmdr2/stable-diffusion-ui is pretty popular, and is a 1-click installer for Win and Linux (Mac coming soon). Quite a lot of features, and well-liked by users for its easy-to-install and user-friendly GUI.