top of page
Writer's pictureAlyxsis
Jun 22

Summer Roundup: AI Workshop Overhaul, Virtual Try On, Groq Playground, HornyHub Updates, Image Editor, Meta's Llama 3, More

Salutations, sissies! Today we have a number of updates deploying site-wide, as well as an exciting overhaul of the AI Workshop, Virtual Try On, an all new Groq Playground hosting 4 cutting-edge models, several new HornyHub models, and more! Let's dig in ~


 

⚒️ AI Workshop Overhaul


The AI Workshop is always evolving but recently has been sorely needing some love. Today we're retiring some of the older models and also reconsidering the types of apps we provide in the Workshop.



👗 Virtual Try On

This has been long in the making, presenting Virtual Try On! Simply upload an image of a person and a garment and watch the magic happen. VTO will automatically generate a mask (the area of the human it is erasing), then recreate the image, inserting the article of clothing on the subject.


You can manually draw the mask if you choose, and optionally enter a description of the clothes for a more accurate fit. Now you can try before you buy even from home!


A sequence of four images showing the process of virtual try-on AI. The first image displays a woman wearing a sleeveless peach blouse and black pants. The second image shows a blue velvet wrap-style blouse with short sleeves. The third image is a masked output with a grey placeholder overlaying the blouse on the model. The fourth image displays the final result with the woman now virtually wearing the blue velvet wrap-style blouse.
A sequence of four images showing the process of virtual try-on AI. The first image displays a woman wearing an orange sleeveless top and light blue jeans. The second image shows a red short-sleeve round-neck t-shirt with a side tie. The third image is a masked output with a grey placeholder overlaying the t-shirt on the model. The fourth image displays the final result with the woman now virtually wearing the red short-sleeve round-neck t-shirt with light blue jeans.
A sequence of four images showing the process of virtual try-on AI. The first image displays a woman wearing a purple and black Vans t-shirt and light blue jeans. The second image shows a white sleeveless button-down top with frill details. The third image is a masked output with a grey placeholder overlaying the top on the model. The fourth image displays the final result with the woman now virtually wearing the white sleeveless button-down top with frill details and light blue jeans.
A sequence of four images showing the process of virtual try-on AI. The first image displays a man wearing a red t-shirt with a blue jacket and light blue jeans, standing against a dark background with plants. The second image shows a white sleeveless button-down top with frill details. The third image is a masked output with a grey placeholder overlaying the top on the model. The fourth image displays the final result with the man now virtually wearing the white sleeveless button-down top with frill details and light blue jeans.
A sequence of four images showing the process of virtual try-on AI. The first image displays a man wearing a black leather jacket, black pants, and glasses, sitting on a stool against a white background. The second image shows a white sleeveless button-down top with frill details. The third image is a masked output with a grey placeholder overlaying the top on the model. The fourth image displays the final result with the man now virtually wearing the white sleeveless button-down top with frill details and black pants, sitting on a stool against a white background.


 
Groq AI Logo.

🚀 Groq Playground

The all new Groq Playground offers access to cutting edge models, faster than you thought possible. Groq's unique LPUs (Language Processing Units) are the fastest LLM inference engine in the world. Currently the Groq Playground offers access to four models:


Meta's Llama 3 70B

Meta's Llama 3 8B

Google's Gemma 7B

Mistral's Mixtral 8x7B

A screenshot of the Groq Playground interface for interacting with open-source LLMs. The selected model is 'Llama 3 70B' with a max token setting of 8192. The user query at the top asks, 'What are some subtle, practical ways I can feminize myself when I go out?' Below, the AI response suggests several methods, including adding a touch of color, dressing for your body type, using accessories, and trying a flowy top. The text is structured with headings and numbered lists for clarity.


 
Meta AI logo.

🦙 Llama 3 & New Standard UI Theme


Meta's Llama-3-8B has also been added to the AI Workshop as a standalone app, like how previous models were implemented with advanced settings.

On a more minor note, we are also debuting a new theme for all of our apps built using Gradio (includes most web apps in the AI Workshop, HornyHub, and the TTS Console). We hope you like the theme; I think it provides a much more cohesive experience across the site when everything matches 💅

A screenshot of the Meta Llama 3 interface. The interface includes a large text input area at the top with the prompt 'Meta Llama 3: Ask me anything...' Below, there are buttons for 'Retry,' 'Undo,' 'Clear,' and 'Submit.' Example prompts are provided, such as 'How to set up a human base on Mars? Give a short answer.' and 'Explain the theory of relativity to me like I'm 8 years old.' At the bottom, there are adjustable parameters for 'Temperature' and 'Max new tokens,' with the current temperature set to 0.95 and max new tokens set to 512.
A screenshot of a text-to-speech (TTS) interface. The interface is divided into two sections. On the left side, there is an 'OpenAI API Key' field, a 'Model' dropdown set to 'tts-1-hd,' a 'Voice' dropdown set to 'nova,' an 'Input Text' field, and a 'Run' button highlighted in pink. Below the button is an area labeled 'Speech Output.' On the right side, there is an 'ElevenLabs API Key' field, a 'Voice' dropdown set to 'Rachel,' an 'Input Text' field with multilingual greetings, and a 'Run' button highlighted in pink. Below the button is an area labeled 'Speech Output.'


 

New uncensored AI Image generator models available in HornyHub

🎨 HornyHub Updates



⚙️ General Improvements

  • HornyHub now stretches to the max width of your screen, to show off your beautiful creations.

  • Much smoother scrolling (Mobile).

  • Smaller UI elements/buttons on most screens (Mobile).

  • Better error messages, should always show up and now easier to read on mobile.

  • Removed a handful of lesser used or old/buggy models.



🖌️ Image Editor

HornyHub now sports a simple image editor! Small but mighty, this nifty toolbar lives in a tab along the top of HornyHub, and allows you to accomplish basic editing tasks to images, including tools to draw, crop, and create layers on your images.

Image generated by HornyHub (Model: Fluently XL Final)

 

🌌 New Image Generation Models


Fluently XL Final

  • Fully uncensored

  • Excellent text/typography

  • Correct anatomy

  • Art and realism in one

  • Controling contrast

  • Great nature

  • Great faces without AfterDetailer

A collage showcasing the capabilities of the new Fluently XL Final image generation model. The images include: A man with a long grey beard and mustache, wearing traditional samurai armor, with a futuristic city background. The rear view of a sleek red sports car with an aggressive design. A police officer in uniform standing in front of a city building. A red-haired woman in a sparkly black dress, with a forest background. A close-up of a leopard's face with yellow flowers in the background. A blonde woman wearing a white crop top that says 'FINAL.' A bear wearing a leather jacket, riding a motorcycle. A young woman in a white hoodie with an artistic watercolor background.

NSFW XL

  • Fully uncensored

  • Removed watermarks

  • Wide variety of styles and mediums


Animagine XL 3.1

  • Fully uncensored

  • Enhcnaced hand anatomy

  • Improved concept understanding

  • Advanced prompt interpretation

  • Fine-tuned from Animagine XL 3.0

A collage showcasing the capabilities of the Animagine XL 3.1 image generation model. The images include: An energetic anime character with blue hair and a blue outfit, surrounded by magical energy. An anime girl with pink hair playing an acoustic guitar, appearing focused and emotional. An anime character with blue hair and a suit, holding a gun in a dramatic pose. An anime girl with brown hair in a red futuristic bodysuit, sitting confidently. An anime character with black hair and a wide smile, wearing a red and blue outfit, with a lively outdoor background.

DreamPhotoGasm

  • Fully uncensored

  • Highly tuned for Photorealism

  • epiCPhotoGasm merged with artistic models


Timeless

  • A mix of Timeless Diffusion and FabulousAlpha (which includes fennPhoto and Incredible World 2) to make a model that doesn't rely as much on negative prompts and that delivers even if "timeless style" isn't in the prompt.

A collage showcasing the capabilities of the Timeless image generation model. The images include: A woman with blonde hair wearing a green bikini and long green gloves, posing confidently next to a vintage car. A man resembling Leonardo DiCaprio in military attire, with a background of fire and smoke. A young woman in medieval armor, standing in front of a stone building with a serious expression. A woman with long brown hair wearing pink sunglasses, a pink bikini, and heart-shaped accessories, standing at an outdoor festival with tents and a crowd in the background.

Pixel Art Redmond

  • This one is safe for you betas out there!

  • Based on SDXL 1.0 and fine tuned on a large dataset

  • Created to be the ultimate pixel art model


Analog Redmond

  • Based on SD XL 1.0 and fine-tuned on a large dataset of analog photographs.


Epic Diffusion

  • This is a heavily calibrated model created by merging dozens of other SDXL models, then blended and reblended multiple times for quality & consistency.




 

💞 Our Sister Servers


I'd like to shine a spotlight on two Discord communities that are aligned with our interests at Alyxsissy, Sissy Positive and the Sissy Research Institute. We are looking to build similar communities and welcome sissy content creators to join & network! Both servers are only 18+.



Sissy Research Institute Discord logo.

Sissy Research Institute

INTELLECTUAL SISSIES NEEDED: ARE YOU A SISSY OR SISSY-ADJACENT INDIVIDUAL INTERESTED IN ANY OF THE FOLLOWING TOPICS:


COMPUTER SCIENCE

LINGUISTICS

BIOLOGY & MEDICINE

DEMOGRAPHICS

PSYCHOLOGY

CONTENT DEVELOPMENT

WORLDBUILDING


IF SO, YOUR EXPERTISE MAY BE USEFUL TO THE SISSY RESEARCH INSTITUTE.



 

Sissy Positive Discord Banner

Sissy Positive

Created by talented hypno creator Sissy Positive (Sophie Rivers), this server runs with the intention to provide a safe and fun place for people within the sissification community to connect and get access to free hot content and interactive features. It's an online community for people to discuss with other like-minded individuals which might help all of us understand ourselves in the context of sissification or kink related interested.



Stay safe, stay sissy,

Alyxsis🖤


0 comments

コメント

bottom of page