<h3 id="latentspace-special">Latentspace Special</h3>What happens if you give first yourself, then Qwen(LLM) and after that Krea2 the fully freedom to be creative? In Case, a AI in Freerun talking to AI in Freerun....Startprompt from me to Qwen:just background,no characters,maybe a very cute animal,illustration,drawing,describe the drawing style,but try a little slightly mixture between an drawing and realistic,a crazy and surreal world,unexpected situations,known things but in a very wrong context,be creative much as you can. Qwen generate from that:In the surreal background, a whimsical yet slightly disturbing world unfolds. Strange flora twists and turns in unnatural forms, their vibrant colors bleeding into one another in unexpected patterns. The air is thick with an ethereal mist that shimmers in a rainbow of iridescent hues.A tiny, cute creature, akin to a cross between a penguin and a koala, navigates this bizarre landscape. Its fur is an iridescent rainbow, shimmering in the light as it moves. The penguin-koala creature's eyes, large and round, are a mismatched pair of bright pink and deep violet, reflecting the chaotic surroundings.The creature's movements are a blend of innocence and awkwardness, as if it is both unaware of the madness around it and slightly amused by it. It carries a small, broken piece of machinery, perhaps a toy or a tool, which seems out of place in this fantastical yet unsettling setting.The overall drawing style combines elements of both realistic and cartoonish illustration techniques, creating a unique visual experience. The details are sharp and crisp in some areas, while other parts are intentionally over-exaggerated and cartoony. The use of bright, clashing colors and exaggerated expressions give the scene a playful yet eerie feel, inviting the viewer to delve deeper into the story or the logic of this surreal world.The scene itself appears to be a playful exploration of the juxtaposition between the familiar and the unknown, with everyday objects and creatures placed in bizarre, distorted contexts. The result is a delightful yet unsettling mix of whimsy and surrealism, inviting the viewer to ponder the meaning behind this bizarre world.And Krea2 gen it out - with no other Rules or LoRA´s!You see the Images - forgive me please &lt;3<h3 id="presetmodel-for-promptless-generation">Presetmodel for Promptless Generation</h3>Concept Model (ah...yes....what....you know....)Concept Release. With this Model wired to Preset Workflow, you can set a "i dont know" Background or Mainscene. Create without LoRA´s. Just out of Base. Its a Model working with 20 Prompts/Vectors handwritten and used Ollama API inside ComfyUI (Qwen) .Category: Concept/SurrealModel standallone is def. SFW. Mux/Merge this Model only with it self! Otherwise you will consult a Eye-specialist/Doctor!Results depending on the used Basemodel from Krea2 Familiy. Have Fun with that, and success while discover the Posibillities. Workflows are inside the Images below.<span style="background-color:rgb(26, 27, 30);color:rgb(141, 141, 143);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Best you merge (mux) this Preset with last (negative) Loadernode to influence. Try different Searchterms like: /candy/rainbow/cute/animal/roadmap/tale/...or anything..../... or use direct dial via Entry Nr.<span style="background-color:rgb(26, 27, 30);color:rgb(141, 141, 143);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Set neg_influence to Searchterm or let randomize self. (descripted in Workflow)NOTE : This Model can only used with my Workflow and the special Customnodes in ComfyUI. It is a transformed Promptcollection, what makes it possible to shorten Pipeline and Generatingtimes.With this System, i aimed to change the Generatingprocess in a different Way. Mainly Prompts can merged or selected by view on Results. Basemodel are all Krea2 Versions and its weights. This is the first Stage and like i told - experimental. Generating Outputs, you can see on my Profile.Instructions to use, you find on my Article Generatingguide krea2. Including the Workflow and the needed Customnodes.Drop Model in Path: ...\ComfyUI_windows_portable\preset_storageInstructions you find: <a target="_blank" rel="ugc" href="https://civarchive.com/articles/33385/krea2-beyond-prompts-a-modular-preset-architecture-for-generative-image-pipelines">here</a>and the latest Version: <a target="_blank" rel="ugc" href="https://civarchive.com/articles/33906/krea2-deterministic-preset-muxing-v10-nsfw">Deterministic Promptless Generationv1.0</a>From now on, you can Record and reuse your Muxing Gen´s.Note: Samples (Images) are generated with my new Dev-workflow. You can use it, but need to ctrl/b blind out missing nodes. (Ollama API Node - not released jet.) If you wanna learn more about, take here a look around and com your suggestions: <a target="_blank" rel="ugc" href="https://civarchive.com/articles/34174/announcement-krea2-full-ai-workflow">Development Stage 2</a>Be a Part of that. &lt;3

Files with SHA256: 83a3d2160ddfc33d779b94afae12d9bfa1d248f8a055038fd9c84638d8854705