*i tagged the flux model as a schnell by mistake, but its a flux dev, sorry! When I look at construction cranes around the city, I see them as giant dinosaurs and I'm pretty sure they have a secret nightlife because they don't stop reproducing ) FLUX sample prompt: dinoCrane by a lighthouse, high detail photo, a blend between a dinosaur and a construction crane, a construction crane in appearance of a robotic dinosaur, high detail photo, ultra detailed digital illustration, no humans, Street photography depth of field, Moody lighting, HDR filter, captured on a Canon EOS-1D X Mark III SDXL sample prompt: a giant dinoCrane - A blend between a big construction crane and a big dinosaur, in japan tokyo city, made of industrial constrution crane, dinosaur face with sharp Teeth, 4 mechanical legs with wheels, craneOsauros crenx ,dinoCrane concept art, punk hipster club-dinosaur with mechanical spikes, pink and purple and orange psychedelic crane design, standing next to a night club at night, high detail digital illustration, digital painting, comics art, hyper realistic 3d art, best quality, ultra sharp and smooth, octane render, high detail realistic photo, ultra sharp focus, Street photography depth of field, Forest atmosphere, captured on a Canon EOS RP the SDXL version is much older but you can get nice results if you refine your prompt the model used for the images in the sdxl version is JJDoe042's-XL - i love it! <a target="_blank" rel="ugc" href="https://civarchive.com/models/430544/jjdoe042s-xl">JJDoe042's-XL - v1.0 | Stable Diffusion XL Checkpoint | Civitai</a> with latent upscale in comfyui using my sampler costume node that can do a latent upscale twice if you need, if you use it try find the sweet spot for the latent upscale denoise, usually its around 0.3-0.5 <a target="_blank" rel="ugc" href="https://github.com/DesertPixelAi/ComfyUI-Desert-Pixel-Nodes">DesertPixelAi/ComfyUI-Desert-Pixel-Nodes: Desert Pixel utility nodes for comfyui</a> <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8c11104a-9e84-41a3-93bd-c44a36f7c52b/width=525/8c11104a-9e84-41a3-93bd-c44a36f7c52b.jpeg" /> and in all my models you can extract at least the positive prompt from the uploaded images with one of my load image nodes <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/07f0210d-f979-4369-a9c2-6a42075a35af/width=525/07f0210d-f979-4369-a9c2-6a42075a35af.jpeg" />

Files with SHA256: b9549353b4a36ab39c54cd4eaa637df97052132bc39a1853952f6a533cb7f025