>>941273369
What you're talking about was an effort made by multiple people to have a second thread for more refined posts that wouldn't slide so quickly. From day 1 you never understood this and have continually reacted like a child. Hence why trolls attacked you. You're immature.
For people who spend time iterating carefully and making a final product the opportunity for recognition is nonexistent.
I've tried to remedy this issue in the past but it always ended up manifesting as 2 threads and this has never been soluble in the past due to a degree of separation between the 2 groups that meant the least frequented one slid. It's only a solution if people are willing to stick with it and that's not up to any 1 person. What you've done is take this suggestion as an attack and assumed because you are a vulnerable person people have maligned you.
Your issue is personal; you've continually acted irrationally and ignored obvious facts. And obviously, I never doxed anyone and you know that I can prove it.
4chan Search
3 results for "112bc656094e87046fbd2d07873cb53b"
>>938969513
The majority of these issues are best managed on the basis of token inversion and concept understanding. Being able to identify a respective model and sampler's management of these things is something that requires technical knowledge. Without that knowledge everything is trial and error. E.g. you wouldn't try to use a new model of an unknown base with random prompts. You would experiment to understand the base first, then work from there. In reality all models have a known base, so you should be able to determine what data blocks are being used and what weights they have to best use the model.
Older gens using more "rudimentary" models can far exceed the output of modern models if their strengths are being played. Improper use of the best models will still produce slop.
The majority of these issues are best managed on the basis of token inversion and concept understanding. Being able to identify a respective model and sampler's management of these things is something that requires technical knowledge. Without that knowledge everything is trial and error. E.g. you wouldn't try to use a new model of an unknown base with random prompts. You would experiment to understand the base first, then work from there. In reality all models have a known base, so you should be able to determine what data blocks are being used and what weights they have to best use the model.
Older gens using more "rudimentary" models can far exceed the output of modern models if their strengths are being played. Improper use of the best models will still produce slop.