∟⊔⊤∦∣≶ @ luthis @lemmy.nz

Posts

144

Comments

2,236

Joined

2 yr. ago

I think you mean this one

This is the one.

If it uses a pruned model, it would be difficult to give anything better than a percentage based on size and neurons pruned.

If I'm right in my semi-educated guess below, then technically all the training data is recallable to some degree, but it's also practically luck-based without having an almost actually infinite data set of how neuron weightings are increased/decreased based on input.

I think this might be intrinsic. Like, there isn't a way to fix this.

I really want to know how this works. It's not like the training data is sitting there in nicely formatted plain text waiting to be spat out, it's all tangled in the neurons. I can't even begin to conceptualise what is going on here.

Maybe... maybe with each iteration of the word, it loses it's own weighting, until there is nothing left but the raw neurons which start to re-enforce themselves until they reach more coherence. Once there is a single piece like 'phone' that by chance becomes the dominant weighted piece of the output, the 'related' parts are in turn enforced because they are actually tied to that 'phone' neuron.

Anyone else got any ideas?

A ready-to-bake lasagna or two in those tinfoil trays. No need to worry about dinner, and one to freeze for another day.

You can use alternative low-carb high fibre pasta. (Just saw the carbs thing) Or just make a different meal.

It's not another 'thing' to clutter the house, and it takes off a bunch of stress.

Really? That's disappointing.

That's two upvotes for Marcus Aurelius!!

I got...

Hyperphantasia A unique thinking style characterized by an extraordinary ability to visualize, enabling a vivid and immersive inner world.

I can pretty clearly create images in my mind, but I just learned I really suck at doing people. Lake with mountains and trees? Well, what kind of trees? Little ripples in the wind or small waves? Snowy or rocky mountains? Easy as. My sister's face whom I just saw? ... uh... kinda?

Grass before beer, you're in the clear. Beer before grass, you're on your arse.

I comment less times per month than days per month so.. it's not that.

As a protonvpn user on linux (arch) I can say that it is totally fine for me. I am using Gnome though.

I'm so dam good at that challenge.

Pretty sure it was the onions.

Note that there isn't a Linux version of the protondrive app. ... I know! what the fuck right?

Secondly, I would just shove Linux Mint onto a USB and use that as a live distro with persistence for a while, just to get used to things. I'm not a fan of debian(/-based) or apt, but it works.

Oh you're right it does work... well fuck knows what I was doing wrong before.

Yeah this is a backup in case I like, mv file to /dev/sda1 or something.

Not a backup of the files, but a backup of the structure.

China does a lot of stupid things, but their train network is admirable.

Financially or electrically?

I don't know much but what I do know is when a government endorses a secure messaging service, it's definitely not secure.

Yes.. I just realised that X has two meanings now. Wait.. 3.