Well, I just quit Twitter and also plan to use Mastodon more even if my main focus is currently on Bluesky.
| Atproto Account: | @chaotichuman.eurosky.social |
Chaotic Human
- 152 Followers
- 172 Following
- 1.3K Posts
infinitist; aspiring shapeshifter who started transfeminine HRT on 5th August 2024; autistic; she/they; age 33; posts in English and/or German
Ok, I could resolve my problem with Twitter now. The violating tweet surprisingly wasn't one of my "cisgender" tweets but actually that one which honestly could indeed be qualified as "violent speech" I guess:
Welp, it happened. My account on Twitter got "temporarily limited" due to "violent speech" (I still guess due to one of my "cisgender" tweets) and due to "technical failures" after multiple log-in attempts despite always successfully passing the Arkose-challenge I now get an "You have exceeded the number of allowed attempts. Please try again later." error when trying to log in again for hours now and that problem can apparently last for days.
However I only recommend using tiling without refiner because the refiner messes up the coloring of the edges of the texture therefore leading to it no longer being seamless. Here are an example texture without using a refiner and that same texture but this time using a refiner.
To all anti-AI-artists who claim "AI-art is just typing in some prompts.": here is an AI-generated image and the custom #ComfyUI-workflow I used to generate it.
Me: *prompts a non existent series ("Shadows of Alagadda") and episode ("The Hanged Kings Trials") styled like an IMDB-entry that is entirely based around concepts from the SCP-multiverse*
#LLaMa: *writes an incoherent non-existent Doctor Who episode that is basically also a character-crossover with many other franchises except anything SCP-related (ok, the name "Hanged King" is still mentioned in the output but probably only because of being in the episode-part of my prompt)*
Anyway: here are the mask (I created with Blender & Gimp) I used as input for the #controlnet-mlsd-model to test it & three resulting outputs (generated with txt2img not img2img).
I used the controlnet-extension (https://github.com/Mikubill/sd-webui-controlnet) for the automatic1111-webui and Deliberate v1.1 (https://civitai.com/models/4823/deliberate) as #StableDiffusion model.
You can find the controlnet-models under https://huggingface.co/lllyasviel/ControlNet
Just found out it's possible to merge the #InstructPix2Pix- with the #riffusion-model by using the receipe in the image of this post.
And the most interesting part here is the resulting instructPix2Pix-riffusion-model indeed still only outputs spectograms however the results are otherwise not that good (I guess the reason is the GPT3-component of instructPix2Pix was not optimized for spectograms) but it's still interesting that this merger kinda works.
I just tested that new classifier-thing by "Open"AI with that infinity-speech I let #ChatGPT meaning "Open"AI's own program(!) write a while back (https://layer8.space/@ChaoticHuman/109604079216997548) and lol.
Chaotic Human (@[email protected])
Attached: 1 image · Content warning: speech by ChatGPT
lol (und ja, ich war tatsächlich mal Mitglied der #Piratenpartei).