I would imagine that they could fabricate most of the parts for other industrial replicators, but there are probably some components that can’t be replicated. We know that dilithium and latinum can’t be replicated, so there are probably other exotic materials too.
How large is the Unification Church? I thought they were a pretty big organization.
Also, this paragraph is hilarious:
The Unification Church, meanwhile, has claimed that engaging in activities that violate Japan's civil law should not be considered grounds for ordering its dissolution and that the government's questioning of the group is illegal.
I don’t really understand the science behind it, but in my experience I’ve had much more success using basic models for training.
Also, I’ve found that LoRAs are generally much easier and faster to train than embeddings. Is there a reason you’re going for an embedding over a LoRA?
Embeddings should generally be trained on base models to improve compatibility with models derived from the base. For SD 1.5, that means using either regular SD 1.5 or the NovelAI leak. You can sometimes get away with using more “basic” models that don’t have many merges, but that can be tough to gauge.
Writing is not hardware-intensive; a Chromebook would be much cheaper if that’s all she does. What else will she use this laptop for that makes you want a MacBook?
Have you looked at Deadlands? It’s a Western RPG system that uses playing cards. That works great for Deadlands because playing cards fit the setting, but I don’t know if that would apply to a fantasy system. Have you thought about using a Tarot deck instead to be more thematic and offer some interesting opportunities with the trump suit?
A couple dozen devices maybe. I don’t really need dedicated ranges, but it’s nice to know exactly which device I’m looking at just by the IP when reading logs.
I know they exist and vaguely what they do, but I don’t know how to set them up. What’s their advantage over simple DHCP reservations for a small client list?
Not from comicbook.com, but close. Looks like you’re right: just more anti-AI nonsense. I wonder if there was this much vitriol when Photoshop first released.
I use Stable Diffusion daily. I’m vehemently against people spouting nonsensical fear mongering against AI. But I completely agree with the author here: a company using AI-generated images in a published book that they charge money for is despicable. AI should be a tool artists choose to use to enhance their workflow, just like Photoshop and tablets. It cannot and should not replace them entirely.
I had no idea that Hasbro had done this. Have they released a statement trying to justify this, or are they just hoping that nobody will care?
I’ve found that AI generation is good if you have a vague idea of what you want, but it can be frustrating if have something specific in mind. I try to approach generation with that in mind: I’ll plan out the large points of what I want but keep an open mind on the finer details.
If I wanted to generate a more specific image, I would first try to do a sketch in another program and then feed that into ControlNet. I haven’t actually done this though since I’m usually able to get something close enough that I can work with.
I meant the quantity of generated images, not the number of tokens. I rarely go over 50 tokens now. As you said, too many tokens and things start to interact in really odd ways. That’s why I’m not a fan of massive lists of negative tokens either; they are much more efficient as a textual inversion like badhandv4 or Easynegative.
However, I only use txt2img to get the rough composition of an image; most of my work is done in inpainting afterwards. If you’re looking to have good images just from txt2img then sometimes lots of tokens are necessary.
Just like traditional art though, this is all based on individual style. It’s important to use what works best for you.
I don’t bother with prompt enhancers any more. Stable Diffusion isn’t MidJourney; quantity is far more important than quality. I just prompt for what I want and add negative prompts for things that show up that I don’t want. I’ll use textual inversions like badhandv4 if the details look really bad. If the model isn’t understanding at all then I’ll use ControlNet.
I would imagine that they could fabricate most of the parts for other industrial replicators, but there are probably some components that can’t be replicated. We know that dilithium and latinum can’t be replicated, so there are probably other exotic materials too.