Art generated from Stable Diffusion / Flux / QWen
Art generated from models & Loras trained on my hand-drawn art
Video Generation
Ai generated and animated using Wan2.2 & Deforum
"The Eternal Garden Pt. 2" Wan 2.2 & Suno
Undersea
Deforum AI animation (SDXL Deforum, trained LoRA, FILM)
"Little World - Part 2" Fractal Variation
Created using Deforum0.7 in Visions of Chaos/Stable Diffusion - my music - DAW: Cakewalk
"Biomorphic" Ai Animation - Stable Diffusion
original music: Drew Medina - SDXL Deforum, trained LoRA, FILM
Experimental Ai
"3D Gaussian Splatting" new 3D scanning technology
Some "Gaussian Splatting" tests! These are tests using stock video & iphone videos to create 3D models using Ai.
Ai NeRF models generated using NVIDIA's "Instant Neural Graphics Primitives"
Ai 3D NeRF models generated using NVIDIA's "Instant Neural Graphics Primitives", sourced from movie arc shots, drone footage and iphone photos. Generated in Visions of Chaos Machine Learning app.
"Magic Garden" GEN-2 text-to-video Ai generated short
The witches wait to gather "ingredients" for their brews... This is purely generated using Ai text prompt to video in Runway's Gen-2.
Beta testing "Wonder Studio" a new Ai video technology to replace people with characters!
Just a FEW CLICKS to auto replace a person with their premade characters, or your own. These are tests using their characters added to my video clips without any rigging, mattes or really anything.
Experimenting with the generation of A.I. art using only a gaming eye tracker
I was considering the potential benefits of A.I. art generation for individuals with disabilities who may have limitations in creating art. Utilizing a Tobii Eye Tracker 5, which I normally use for gaming, I was able to fully control the mouse and generate basic A.I. imagery independently.
Tools and Techniques
I try to stay Open Source, compute costs are crazy currently on the best models.
Main software: "Visions of Chaos" software, colab notebooks
- Image models: Stable Diffusion SDXL, Flux, Fooocus
- Model training: Trained checkpoint models & Loras using my own original art, to create new original art in my style.
- Video creation: Deforum, FIFO, Stable Video Diffusion, Animate Diff with Prompt Travel,
- Outpainting and Inpainting diffusion (extending art and editing using Ai)
- Depth mapping: Depth map creation for games, sculpture, lenticulars
- 3D mesh creation: Splats/Nerfs/scanning/2D to 3D mesh tools
- Audio: Stable audio, Custom Ai TTS voices
Copyright © 2024 Drew Medina - All Rights Reserved.