liliumstar

Air is actually good, but they don't have a lot of fast servers. You are naturally limited by the server you choose and peering.

MakeMKV is non-free proprietary software. It just happens to be free while in beta, which it has been forever. There's not a lot of great free software solutions that do the same thing, in fact it's the main (or only) way people extract 4k BDs with the FEL intact.

I know you said consumer GPU, but I run a used Tesla P40. It has 24 GB of vram. The price has gone up since I got it a couple years ago, there might be better options in the same price category. Still, it's going to be cheaper than a modern full fat consumer gpu, with a reasonable performance hit.

My use case is text generation, chat kind of things. In most cases, the inference is more than fast enough, but it can get slow when swapping out large context lengths.

Mostly I run quantized 8-20B models with the sweet spot being around 12. For specialized use cases outside of general language, you can run more compact models. The general output is quite good, and I would have never had thought it was possible 10 years ago.

ETA: I paid about $200 USD for the P40 a couple years ago, plus the price for a fan and 3d printed shroud.

I would do FDE yeah. My current laptop setup is with systemd-boot and a special initramfs that allows me to unlock it with a yubikey, with fallback to password. Fair warning, this exact configuration is not particularly easy to setup.

There are also modules which enable early network connectivity along with a SSH server, meaning you login and unlock it remotely. I have not tried this.

Debian does not frequently require rebooting under normal circumstances. Kernel updates are not that frequent, and you can usually put it off for a bit if you don't want to deal with it.

CRF 18 is usually where I start at, but that is focusing on quality. I'd imagine you would want a higher starting point.

If you want to find out what different options are or do, the docs are a good place to start: https://x265.readthedocs.io/en/master/cli.html

In my opinion the ideal x265 size/speed/quality is using a tuned slow preset, perhaps with filtering if the source is grainy. A test encode or few should be done to determine an ideal CRF per source.

Since you don't seem very familiar with x265, I would just stick with the defaults in slow preset, but consider using aq-mode 3 or 5 (only available in the patman mod). You can also adjust the aq-strength to help control the resulting size somewhat, I wouldn't go lower than 0.5.

I tried to replicate this in KDE (wayland).

pynput has no effect on wayland, which I figured. I then enabled ibus wayland to type unicode points with ctrl+shift+u, but everything came out as expected. So I think it is related to the xorg backend of pynput.

Yup, I think it'd work fine, especially if you want the ability to easily inspect individual items.

Any of the popular python yaml libraries will be more than sufficient. With a bit of work, you can marshal the input (when reading files back) into python (data)classes, making it easy to work with.

I would scrape them into individual json files with more info than you think you need, just for the sake of simplicity. Once you have them all, then you can work out an ideal storage solution, probably some kind of SQL DB. Once that is done, you could turn the json files into a .tar.zst and archive it, or just delete them if you are confident in processed representation.

Source: I completed a similar but much larger story site archive and found this to be the easiest way.

I'm with Azire, they have port forwarding and 10 gig servers. Note they were bought recently by malwarebytes, so it is possible things will change in the future. For the time being, things have been great. I moved from OVPN after myself and others started experiencing persistant failures.

I've been meaning to try out CryptoStorm. If anyone has experience with them please share.

Congrats! I just got a similar running on Arch with a 5700 XT. When I looked at it a couple years ago, it wasn't really possible. Now, smooth sailing.