Jiayu;  "Jiayu;" is a pivotal piece in this AI series, built upon the foundation of  LoRa model training and generative techniques, establishing a creative process that fuses art with artificial intelligence. This work employs advanced AI models trained on digital images from the artist's previous works, drawing inspiration to instruct the AI to understand the artist's unique visual language and construct its worldview. Components from the artist's past works-such as particle effects, fluid dynamics, three-dimensional scene construction, composite visuals, and camera movements-have been abstracted into keywords to train the AI. Through this process, the AI has learned to conceive a natural world composed of abstract particles, encapsulating symbolic elements the artist has focused on over the years, including mountains, streams, forests, grass, mushrooms, flowers, white birds, and bamboo groves. Once the AI has completed its learning and begins to generate outputs, it creates a new visual world in response to prompt words provided by the audience. Each audience interaction infuses the work with a unique personality as if heralding the "aura" of art in the technological age-the fleeting, elusive charm of art-rekindled through participatory creation. The dynamically generated imagery results from the interplay between art and technology and challenges the traditional physical and temporal constraints of artworks, seeking to carve out new dimensions in the art experience. Through this dynamic creation process, the series reflects on how art's uniqueness and existential value can be preserved and prolonged in an era where digital reproduction and dissemination are increasingly accessible. This is not just a reinterpretation of the concept of the "Aura" in the age of technology but also a profound exploration of contemporary art practice and theory. Prompt：A trigger word is (made of particles), and by setting the weights between 0.65-0.8,you can get a better result :)Let's quest Aura together ~!

Jiayu - Art Landscape Particles(Beta)

beta-AllForPraticles.safetensors

v1.0

This is a LoRa model that trained by sequence frames of our new_media past works, and it's the Beta version.This model can generate our Particle-style scenes, especially forests and rivers.Our goal is training a Lora model to be able to convert objects and scenes as particles or point clouds.Trigger word is (made of particles), and setting the weights between 0.65-0.8, you can get a better resultcheckpoint models recommend: MajicmixRealistic, Cyberrealistic, Onlyrealistic

Files with SHA256: c644fa63aecc231bc036867332bd27cc97acad3aa0f2bee3d89d7fea26c8e998