80085
I just looked it up, and apparently someone implemented dynamic activation functions in a CNN: https://www.nature.com/articles/s41598-022-19020-y . I’ve never seen something like this elsewhere. I have included various activation functions in hyperparameter searches before full training to find the “best” one on datasets. I haven’t really seen much of a difference in validation performance between activation functions.
Found another paper using dynamic activation functions with transformers: https://arxiv.org/pdf/2208.14111.pdf
“If Rome possessed the power to feed everyone amply at no greater cost than that of Caesar’s own table, the people would sweep Caesar violently away if anyone were left to starve.”
- Eben Moglen
I think imposing artificial scarcity on art, information, and tools; and rationing based on those with the ability to pay is immoral. I mean sure, most art that people pirate is just empty entertainment. But imposing artificial scarcity on tools (software such as OSs, CAD, productivity software, etc), news, and academic papers behind expensive licenses that many cannot afford to pay is objectively immoral. If piracy did not exist, I am positive the world would be without many of the technological advances we have today.
Is he really a hopeless candidate? Misogynist-sociopaths seem to have been making a comeback lately. Trump’s chances in the primaries would be weakened if convicted. Biden’s health hasn’t gotten any better, and he has always been a weak, uninspiring politician.
I’m not a crypto supporter. I do find the tech a bit interesting. I guess, tech-wise, Lemmy would be more comparable federated crypto like Stellar or Ripple (dunno if these are still federated, or if more popular federated crypto exists; been a very long time since I kept up with it). Without some sort of “trust” decentralized systems are too expensive (resource-wise) to be worth it, IMO.
Off-topic, but I’m kinda surprised p2p networks haven’t really advanced since Gnutella. I believe they had the concepts of trust/reputation and self-organizing networks with “super-peers” way back then.
Nobody really knows because it’s an OpenAI trade secret (they’re not very “open”). Normally, it’s a hard limit for LLMs, but many believe OpenAI are using some tricks to increase the effective context limit. I.e. some people believe instead of feeding back the whole conversation, they have GPT create create a shorter summaries of parts of the conversation, then feed the summaries back in.
Threads doesn’t need to do an EEE attack. They’ve already gained many more users than the entire Fediverse. At this point, I wouldn’t be surprised if they decided to not join the Fediverse at all.
I would never use Threads, but I would use a Mastodon instance that federated with Threads. I already see many journalists and content creators I like trying it out, who either stopped using Mastadon long ago or never even tried it in the first place. If Threads started doing things that negatively affected my experience, I would then switch to a Mastodon instance that wasn’t federated with Threads.
I’m not too well-versed in world-history, but when you look up definitions of “left-wing,” and the policies Stalin implemented, there doesn’t seem to be much overlap. Seems like a lot of authoritarian leftists just used the idea of a vanguard party to seize power and implement policies closer to fascism and state-capitalism than those of communism. Similar to the National “Socialist” Party.
