Using text prompts to generate specific shapes and designs is and still will be quiet a challenge. With the help of image based guide maps, this challenge becomes quiet enjoyable!
This is a first attempt to create a custom fancy font generator. Enter some text and the code will use alternative Unicode letters.
This experimental approach explores the co-creational potential of working with generative ai on plain color /drawing based interaction. In addition to the visual creation, we use LLM’s and proper systemprompts in combination with audio generator models to use the generative graphics as a musical score.
Upscaling is a crutial feature when working with generative tools. The range begins with simple upscaling and sharpening specific details. In more advanced methods, the upscaling process can modify the original image in specific directions using prompts and specific LoRas.
This is a kickoff of a series of webbased prototypes, that target to use visual material as a basis for …
With the LTX Video 13b distilled model, we have a resh opportunity to generate short animations in quiet quick and responsive way. The 13b Model has quiet comparable prompt understanding and with current start- and endframe input we have proper control over the diffusion process.
ComyUI has, besides its generally complex workflows, the capability to offer basic image manipulation with GenAI. This comes with basic …
This experiment is a simple comparison between both Depth based diffusion pipelines. The older, pretty robust ControlNet Depth with Stable …
This is a row of user interface experiments that uses large language models and simple system prompting to generate drawing …
DOWNLOAD WORKFLOW low_step_upscale
Systemprompts sind neben den üblichen Prompts, übergeordnete Anweisungen speziell für Sprachmodelle. Wohlgeformte Systemprompts haben maßgeblichen Einfluß auf den Output des …
simple generative object made from primitives
First off, this tutorial is based on designers trial and error best practice knowledge using Kohya, basic knowledge and several …
act driven approach The most direct and relatable approach is by using an acting video, that drives both visual and …
basic dataset The starting point of this experiment is a dataset of pretty clear images of black and white textile …
Working with layouts and typography based designs is quiet challenging with the current tools available. As typography comes with very …
With LLM’s we basic text enhancement methods image remix methods We take a generated image with a relative simple prompt. …
The creation of images or even videos with the approach to synthesize authenticity is quiet challenging in many ways. Besides …