ragnarokonline

@ragnarokonline@vlemmy.net

This profile is from a federated server and may be incomplete. Browse more on the original instance.

ragnarokonline,

I’m pretty tired and my dang dad wont stop calling me to help him set up his new phone. But imma make it through. Thanks for askin’ homie.

ragnarokonline,

One foot in front of the other, homie. All you have to do is focus on making it past the next 10 minutes in front of you.

You come from warriors, and their strength lives in you.

ragnarokonline,

Wish I could help ya. You definitely want a hardened finish like poly or shellac, but shellac is going to stink up your home if you can’t remove the banisters to finish.

I’ve had a lot of luck with beeswax for a finish, but more for boxes than furniture. There are a lot of great bands out there, though you’d probably have to reapply every 6-12 months (which would get tedious)

ragnarokonline,

For some people, it’s the opposite: when they get a partner, their relationships with friends weaken.

Having someone around to help take care of you whenever you’re at end-of-life is definitely a perk of having a partner, though.

ragnarokonline,

Thank you for sharing your journey so others can learn. Community is our only defense against the robots.

ragnarokonline,

Just switched and it looks great! Thanks to everyone got their excellence and major hustle in getting this out.

ragnarokonline,

It's been one week since spez looked at me Cocked his head to the side and said, "I'm angry" Five days since I joined Lemmy Saying, "Get your API together, come back and see me" Three days since the living room I realized it's the IPO’s fault, but couldn't see it through Yesterday, Spez was blocking me But it'll still be two days 'til he says he’s sorry

ragnarokonline,

I’m sure you can find a community that will be cool with you posting whatever you’d like.

That said, it’ll probably be a community filled with like-minded people, so don’t plan on converting many folks to your ideas. You’ll probably be surrounded by people who are likely to agree with your perspective anyway.

ragnarokonline,

Probably not what you’re looking for (it’s EDM; after the buildup there’s a drop, etc…), but I tend to call Turbulence by Steve Aoki “The buildup song”

https://youtu.be/y3OzHBEcymw

ragnarokonline,

Open em up and get those wonderful lil magnets out to add to your magnet collection

ragnarokonline,

This is so true for me too. Mlem will always be a part of my journey away from Reddit and towards a more open mind.

We need to have a way to subscribe to communities when someone posts a link to it on c/newcommunities or in any comment/ post.

Right now the only way to subscribe to communities is through the desktop website. It would make using the app, as well as the platform far more intuitive if we could click on a link (with whatever format is acceptable) that doesn’t just open the community in the in-app browser, but actually took you to the community from your...

ragnarokonline,

Agreed. Love Mlem, and am looking forward to this. As sharing communities gets easier, I’ll be excited to see more things like communities on one instance linking to communities on another server that have similar subject matter.

andrewginns, to machinelearning
andrewginns avatar

It's been a personal mission to get up to date on Transformers and derivative architectures given their outsized impact. Attention is the key to this and today I came across the most intuitive explanation I've found so far from George Hotz.

  1. In a traditional NN we have a fixed set of weights for the model that do not change based on the input
  2. Now imagine we could change the weights in a NN based on the context of the problem
  3. For example we could have 256 weight sets that can be used for each layer, we could then choose which of the weights to load in based on the context of the data
  4. Applies to both training and inference data contexts

This is essentially what Attention in Transformers does. Rather than having 256 different sets of weights to swap out we have a continuous representation that provides the context and the weights needed for the model to adapt to that context. This is the ‘trick’ to Attention.

Adapted and expanded on it after listening to the Latent Space Podcast Ep 18.

ragnarokonline,

Nice explanation, but as someone who’s still just learning ML, it’s amazing to me that the weights can dynamically change depending on the challenge. Especially after so much training is done to set the weights in the first place.

I may not be fully understanding still, though.

ragnarokonline,

Earth.nullschool is my go-to when Hurricane season rolls around for my relatives.

ragnarokonline,

Nice to be reminded that the old internet still exists

  • All
  • Subscribed
  • Moderated
  • Favorites
  • megavids
  • kavyap
  • DreamBathrooms
  • thenastyranch
  • magazineikmin
  • InstantRegret
  • GTA5RPClips
  • Youngstown
  • everett
  • slotface
  • rosin
  • osvaldo12
  • mdbf
  • ngwrru68w68
  • JUstTest
  • cubers
  • modclub
  • normalnudes
  • tester
  • khanakhh
  • Durango
  • ethstaker
  • tacticalgear
  • Leos
  • provamag3
  • anitta
  • cisconetworking
  • lostlight
  • All magazines