The Experimental Zone: AI, VR, AR, Code Art

Custom AI Model: The Photographer

Since late 2022, I have been focused on developing a custom AI model for use in Stable Diffusion or any platform that supports 'safetensors' files. The model, called The_Photographer, has been designed to produce photo-realistic images of a diverse range of models. The goal is not simply to create photorealistic images, but to replicate the framing, composition, and style that a professional photographer would create when capturing an image with a camera.

So far, the model has the capability to produce a wide range of ethnicities, body types, clothing ensembles, backgrounds, facial features, natural-like imperfections, and physical poses. Additionally, The_Photographer can imitate various photo styles, such as old film, monochrome film, digital, depth of field, bokeh, wide shots, close-ups, low light, high key light, moody lighting, and shadowy elements.

ThePhotographer_v1.00 was made with a mixture of custom training and checkpoint mixing.

  • Images come out best in 512 x 512 but can produce really good results at resolutions exceeding 712 x 712.

  • Use the sampler DPM2 Karras for more photo-realistic results.

  • You can use popular camera and film types as descriptions for corresponding results.

  • It's capable of producing nsfw results.

  • Released under the CC-SA 2.0 license.

ThePhotographer_v1.00 is available for download here:

NOTE: By downloading and using this AI model, you assume full responsibility for all content generated and its potential consequences. The creators of the model are not liable for any misuse or resulting harm.

Images generated with ThePhotographer_v1 with no additional post-editing.

ThePhotographer_v1.20 is an improved version that understands a wider range of styles, poses, and locations.

  • Tbh I've been rendering images at all resolutions without an issue, so 512x512 is not necessary.

  • Use the sampler DPM++ 2m for more photo-realistic results at a good rendering speed.

  • You can use popular camera and film types, different eras of photography, different angles, and different locations as descriptions for corresponding results.

  • Uncensored and capable of producing nsfw results. Use Responsibly. Warning: May produce likenesses of celebrities and public figures.

  • Released under the CC-SA 2.0 license.

ThePhotographer_v1.20 is available for download here:

NOTE: By downloading and using this AI model, you assume full responsibility for all content generated and its potential consequences. The creators of the model are not liable for any misuse or resulting harm.

Gonzo Illustrator

This model is designed to produce American Style 2D Illustration & Comic Book aesthetics

Gonzo Illustrator is a custom-built AI model crafted for artists, illustrators, and comic enthusiasts seeking American-style 2D illustrations, classic comic book styles, and expressive character creation. Built on Stable Diffusion 1.5, this model combines custom training with checkpoint mixing to bring you a versatile tool for creating vibrant and detailed illustrations.

Features

  • American Comic Book Style: Recreates the distinct, bold style of American comics, perfect for both classic and modern visual storytelling.
  • Detailed Character Creation: Optimized for creating expressive characters with depth, enabling artists to build visually compelling heroes, villains, and everything in between.
  • Versatile Styles: With the ability to mimic popular 2D illustration styles, Gonzo Illustrator is adaptable for various themes, from light-hearted humor to more intense and mature artwork.
  • NSFW Capabilities: This model includes the flexibility to generate both SFW and NSFW content, offering creative freedom to artists.
  • Open-Source Licensing: Released under the Creative Commons Attribution-ShareAlike 4.0 (CC-SA 4.0) license, making it free for adaptation and sharing, provided credit is attributed.

Gonzo Illustrator is available for download here:

NOTE: By downloading and using this AI model, you assume full responsibility for all content generated and its potential consequences. The creators of the model are not liable for any misuse or resulting harm.

DreamPlace_ArchiGen

This model has been designed to produce images of architecture, landscapes, and interiors.

ArchiGen is an AI image generation model designed primarily for producing architecture, landscapes, nature scenes, backgrounds, and experimental structures. It can also create various urban elements and simulate different weather conditions. This model is compatible with Stable Diffusion or any other generator that accepts .safetensor files.

DreamPlace-ArchiGen is available for download here:

NOTE: By downloading and using this AI model, you assume full responsibility for all content generated and its potential consequences. The creators of the model are not liable for any misuse or resulting harm.

Brutalis: Brutalism in a Box (LoRA)

Brutalis is a specialized LoRA file focused on adding brutalist style and elements to your architectural, structural, or artistic renderings. Drawing from various brutalist styles in Eastern and Western regions, as well as distinctive structures and sculptures from countries like Russia, Germany, Africa, and United States locations - this file offers an authentic brutalist touch for your creations.

With Brutalis, you can easily incorporate signature brutalist characteristics into your projects, ranging from architectural designs to building structures or conceptual art. This tools adds an invaluable layer of brutalism to your artistic vision. Experiment with the raw beauty and unique features of brutalism using Brutalis.

This model works great when used along with DreamPlace ArchiGEN.

Trigger words

You should use <lora:Brutalis:1> or Brutalis to trigger the image generation.

To see examples or visit the repo, check out my HuggingFace

A Touch Of Color - ATC (LoRA)

A Touch of Color (ATC) is a LoRA designed to enhance the generation of photorealistic images of women from diverse ethnic backgrounds, including African, African American, Indian, Asian, and mixed-race subjects. By refining supermodel-like features, ATC addresses the limitations some Stable Diffusion (SD) models face in accurately generating brown-skinned individuals and recognizing ethnic tags. Although it is specifically trained on female subjects, ATC delivers high-quality results, ideal for makeup editorial commercial closeups that highlight facial and eye details. Additionally, it excels at rendering tanned white-skinned individuals, complete with tan lines and freckle details. Activate by using the tag ATC or ATouchOfColor:1 inside left and right angle brackets.

Key Features:

  • Compatibility: Works seamlessly with Stable Diffusion 1.5 models.
  • Resolution: Capable of generating images in various resolutions.
  • Tag Recognition: Accurately understands and responds to a wide range of tags, including:
    • Editorial
    • Dark skin
    • Brown skin
    • African
    • African American
    • European
    • White skin
    • Tan skin
    • Asian
    • Japanese
    • Korean
    • Indian
    • Mixed race
    • Close up
    • Glamour
    • Makeup editorial
    • Boho
    • Bohemian
    • Bokeh
    • Depth of field
    • Full Lips
  • NSFW Capabilities: Capable of generating NSFW content.

Trigger words

You should use ATC or <ATouchofColor:1> to trigger the image generation.

To see examples or visit the repo, check out my HuggingFace

Continued AI Research

I'm absolutely in love with the new AI technology hitting the digital world and spend many days and nights experimenting with it. As an artist I don't feel threatened by AI at all, I actually feel creatively enriched by it. I don't have access to a lot of extra capitol and can't afford to hire a studio full of people to work around the clock and help me create all of my weird ideas, but with some of these AI tools I can level up my individual productivity.

Currently I'm experimenting with:

  • Training AI Image Generation Models
  • Training AI Voice Manipulation Models
  • Exploring AI Text Generation
  • Exploring AI Sound Generation
  • Exploring AI Audio Manipulation
  • Building Automated AI Tools
  • Using AI to replace certain steps in my workflow

Metaverse Experiment: Virtual Art Galleries

Over the pandemic during the whole "metaverse" craze I experimented heavily with blockchain technology and several metaverse-friendly creative tools. My plan was to build somewhat of a virtual amusement park with a variety of galleries, experiences, virtual shops, and games. So far I've created about 4 virtual galleries and 1 little metaverse experience for a company in Europe. 2 out of 5 experiences are still available online. The other 3 either started to break because of blockchain platform errors or disappeared because the company hosting it failed.

 

I'm still experimenting with this idea and am working towards building more virtual spaces but don't expect anything to pop up in 2023, unless I get a burst of inspiration or a custom commission.

BLKMRE Voxels

TWO-D Token Vault

The Temple Void

Metaverse Experiments

Space Description Availability
The Temple Void A Virtual Gallery displaying NFT Artworks hosted on Hic et nunc, a platform hosted on Tezos. Available for Mobile, Desktop, and VR CLOSED due to HEN closure
TWO-D Token Vault A special vault containing Blackmoore's first collection of hand-drawn NFTs minted to various blockchains. Available for Mobile, Desktop, and VR CLOSED due to Mozilla Hubs Discontinuation.
BLKMRE Voxels Gallery A gallery hosted on the Ethereum Network highlighting various high-end NFTs by Blackmoore. This space is located in the ONE/OFF district. Available for Mobile, Desktop, and VR OPEN
Socialogue Metaverse Experience A side metaverse experience that users of the blockchain platform Socialogue could explore. CLOSED due to Socialogue closure

UPDATES:

So Mozilla Hubs shut down recently and that caused the closure of my TWO-D Token Vault. It's interesting to see how many of these webVR projects don't stand the test of time.

Alternate Reality reactive Physical Art

I see myself as a 'phygital' artist, meaning I love diving into both physical and digital art forms. Mixing the two is what I like to do! Whether it's merging a physical object with a digital twist, like a cassette tape with electronically made tunes, printing out digital art, or painting something that comes to life digitally through a phone scan, I'm all about that fusion.

Over the last few years I have been experimenting with AR reactive prints and paintings as a way to add more interactivity to art events and exhibitions. I typically use 3rd party apps to host my AR experiences but I'm in the process of creating my own AR viewer that will be accessible from this website.

Examples

In this video I show a few examples of AR reactive art that I've made using 3rd party apps and tools like Meta's Spark Studio.

 

  • The first clip shows the AR layer to a poster I made of Vincent Van Gogh in a straight jacket during my residency at Lighthouse Immersive for the Immersive Van Gogh exhibit. This is a "pop-up" style animation that happens when the poster is scanned. This one is currently available on my Instagram in the filters.
  • The second clip is another pop-up style action that happens when my Paulie Walnuts poster is scanned. This one has an audio layer.
  • The third clip shows an animation that happens when my "Red" poster is scanned. This one is also available on by Instagram profile in the filters.
  • The fourth example shows a painting called "Glitch Girl" that has 2 different AR layers: one that shows a glitch animation and another that shows a pop-up with transparent lenses. This piece was featured in the Padzieski Gallery during my Tead One residency. This effect is also available on my instagram profile.
  • The final clip is a pop art inspired poster that has an AR animation layer.