Pygmalion Links
Website • Matrix • Github • Twitter • HuggingFace • Reddit profile • Subreddit • Telegram • Discord • Wiki
📜 The Pygmalion Docs — If you're new to Pygmalion, please read this first. This will answer most of your basic questions.
☁️ To run Pygmalion on the cloud, choose one of the links below and follow the instructions to get started:
- TextGen WebUI — Simple CAI-like interface. Also known as Oobabooga. Runs on a Gradio app, but you can also connect it to SillyTavern for a more feature-rich interface.
- KoboldAI Horde — The Horde lets you run Pygmalion online with the help of generous people who are donating their GPU resources.
- Agnaistic — Free online interface with no registration needed. It runs on the Horde by default so there's no setup needed, but you can also use any of the AI services it accepts as long as you have the respective access key.
🖥️ To run Pygmalion locally (on your own computer), follow one of the guides below:
Make sure your machine meets the system requirements before downloading.
- TextGen WebUI — This guide will help you install the TextGen WebUI (also known as Oobabooga).
- KoboldAI — This guide will help you install KoboldAI, a browser-based frontend.
- KoboldAI 4bit — For GPUs with less than 10GB of VRAM
- koboldcpp — Run KoboldAI on CPU only, using 4-8GB of RAM. No $2000 GPU necessary.
- SillyTavern — Atmospheric interface. This is only a UI; you need to connect it to a service like TextGen, Horde, or KoboldAI.
🤖 To create and share characters, refer to the links below:
- Character Card Creator — Most interfaces use character cards, which are images that have extra data inside them that define the character.
- Character Hub — A place to share and download character cards (the site is SFW by default but has a NSFW toggle)
The following is an aggregation of links, guides, and tools related to Pygmalion and other open-source Large Language Models (LLMs). The field of LLMs is ever-changing. Some of the links below may be outdated, but the author of this Rentry will try to keep it up to date.
Pygmalion Models
Pygmalion 2 and Mythalion
Pygmalion 2 is the successor of the original Pygmalion models used for RP, based on Meta's Llama 2. Mythalion is a merge between Pygmalion 2 and Gryphe's MythoMax.
Pygmalion 2 Models:
Quantized by TheBloke:
- Pygmalion 2 7B GPTQ (GPTQ models use far less GPU resources, at a small precision cost.)
- Pygmalion 2 7B GGUF (GGUF models can run entirely on CPU.)
- Pygmalion 2 13B GPTQ
- Pygmalion 2 13B GGUF
- Mythalion 13B GPTQ
- Mythalion 13B GGUF
Pygmalion 1 and Metharme
Pygmalion 13B and 7B are chat models that use Meta's LLaMA as a base.
Metharme 13B and 7B are experimental variations of Pygmalion 13B/7B fine-tuned for instruction.
NOTE: The 13B / 7B models were released as XOR files due to licensing issues re. Meta, and cannot be used as-is. A list of converted models can be found here:
Older Models
These models are generally outdated and not guaranteed to work on some text generation services.
- Pygmalion 6B — Based on EleutherAI's GPT-J 6B
- Pygmalion 2.7B — Based on EleutherAI's GPT-Neo 2.7B
- Pygmalion 1.3B — Based on EleutherAI's Pythia 1.3B
- Pygmalion 350M — Based on Facebook's OPT 350M
User Interfaces
Colab users: Due to high resource usage, Google no longer allows the string PygmalionAI
in Colab code. You can still use Colab with other models. Almost all Colabs now use different unbanned versions of Pygmalion.
"The Website"
Developers have been working on an official UI, which will eventually be hosted here. Have patience. Remember: they're doing this all for free, in their own spare time.
In the meantime, you can enjoy Pygmalion using one of the following interfaces:
TextGen WebUI
A combined front-end and back-end for running models like LLaMA, GPT-J, Pythia, etc.
- Installation guide by Oobabooga
- Installation guide by PygmalionAI
- Colab — Simple CAI-like interface with a list of optional parameters that you can enable or disable before running. Works on mobile.
- Local Installers — One-click installer. For Windows, Linux, and MacOS systems.
- Github
- Subreddit
TavernAI
An atmospheric interface with lots of customization options, that uses a card-based character file system. Note that TavernAI is only a front-end; it needs to connect with a back-end such as Kobold (see below) or TextGen (see above).
- Original TavernAI Colab — Note: no longer supports Pygmalion-related models.
- Mobile Colab — A fork of the above Colab, with better UI scaling for mobile users, and runs Pygmalion-related models. Uses an older Tavern version.
- SillyTavern — A Tavern fork that adds many more features over the original Tavern.
- Colab
- Note: This Colab only hosts the Extras Extensions API, not the actual Web UI. To use the Web UI, you need to install SillyTavern to your machine first; read this simple guide to learn how.
- SillyTavern Docs
- SillyTavern Subreddit
- Colab
KoboldAI
A back-end for running a wide range of open language models.
- Installation guide (GPU) — For GPUs with 16GB of VRAM or more
- Installation guide (4bit) — For GPUs with less than 10GB of VRAM
- Installation guide (CPU) — No GPU necessary
- TPU Colab — Runs larger models.
- GPU Colab — Runs smaller models.
- KoboldAI Horde Client
- Kaggle — Alternative to Colab, but can be very janky at times.
- koboldcpp — C++ implementation. Use it to run GGML/CPU-only models.
- KoboldAI Github
- United fork of the original KoboldAI page and several other forks. Generally more updated.
- Subreddit
Other UIs
Chat Websites
- Harpy Chat
- Spicy Chat
- Crushchat
- Chub Venus
- Joyland
- Charstar
- Chatfai
- EasyERP
- Persona AI
- Unhinged AI
- OpenCharacters
- Poe
Open-source Platforms
- miku.gg — Open-source UI to run chatbots. Allows visual and text-to-speech (TTS).
- RisuAI — Tavern-like frontend, feature-rich and has plugins. Run in web or with installer.
- AgnAIstic — UI based on Pygmalion's WIP Galatea interface. Runs chatbots anonymously, no login required.
- Faraday.dev — Run open-source LLMs on your computer. Completely offline. Zero configuration.
- Open Playground — An LLM playground you can run on your laptop.
- Laika's LiteVN UI — Highly customizable visual novel-like UI.
- RenAI Chat — Virtual Novel interface to chat with AI models.
- MonikA.I. — "This project aims to add new AI based features to Monika After Story mod with the submod API."
- Web UI with LTM — Experimental Web UI with Long Term Memory (LTM), which helps the bot remember long-term conversations.
Guides / Tips
General Guides
- Official documentation page — Answers most of your basic questions about Pygmalion.
- You are encouraged to contribute to the documentation page here.
- Pygmalion Wiki — Work-in-progress Wiki.
- Alpin's Pygmalion Guide — Very thorough guide for installing and running Pygmalion on all types of machines and systems.
- Spiritual successor to the old rentry guide.
- Local Installation Guide (CPU) — You can run Pygmalion on CPU only, with only 4-8GB of RAM (not VRAM) required, thanks to GGML.
- "Okay, so what the heck is a chatbot?" — Very simple and helpful guide to getting a chatbot up and running on Tavern.
- Soft Prompts
- Read more about soft prompts here
- The Novice's LLM Training Guide — Read this if you want to learn how to train your own LLMs
- CharacterAI chatlog dumping
- CAI data scraping tool
- Upload your CAI dumps here to help train Pygmalion.
- General guide to open-source chatbots
Character Guides
- Bot Creation Guide
- Ali:Chat — "Ali:Chat’s principle idea is using dialogue as the formatting to express and reinforce traits/characteristics. Ali:Chat can either be used by itself or combined with another style."
- Ali:Chat Lite — Token-optimized character format
- PLists — "PLists are where you describe the traits of a character. This includes personality, clothing, body, and even the environment of your RP."
- Pygmalion Tips
- Pygmalion character creation + writing tips
- How to make a character
- Importing CAI characters to Pygmalion
Character Repositories
- Character Hub — Share and download Tavern cards (SFW; has a NSFW toggle)
- Tavern card booru — Share and download Tavern cards (NSFW)
- BotPrompts — List of JSON files for Pygmalion (Inactive)
- /aids/ Prompts — Story prompts for KoboldAI bots
Tools / Scripts
- AVAKson's Character Editor
- JSON character creator — Enter your character info to create Pygmalion-compatible JSON files
- TAIConvert — Convert TavernAI cards to JSON files
- CAI tools — A Chrome extension made for CAI data scraping. Allows you to download all of your CAI bot's chat history, as well as their definitions (if you have access to them).
- More helpful CAI tools and scripts — HYW, fix tildes, darken italics, and more.
- List of CAI scripts on Greasyfork
Additional Open Models and Resources
- Awesome Marketing Datascience — An absolutely exhaustive list of resources and tools regarding LLMs, image/audio generation software, and more.
- The OpenLLMs List — Almost every open LLM you can think of is listed here.
- List of Tools and Resources
- Which Open LLMs are the best?
- TheBloke's model list — Hundreds and hundreds of open models to choose from, take your pick.
- Awesome Totally Open Chatgpt — A list of open-source alternatives to ChatGPT
- Pickaxe Project — Build, manage, and deploy custom LLM tools
- There's An AI For That — Massive list of user-submitted AIs from a wide variety of categories, including conversational, image gen, audio, and a lot more. The site itself was written by ChatGPT.
- Simon Willison's Weblog — News on LLMs and machine learning
- Machine Learning subreddit
- Local LLM links
- Local LLM links
- Local Model General Resource Guide
- EmbedBase — Open-source SDK and API to easily feed data to LLMs
- LaMDA RLHF Pytorch — Open-source pre-training implementation of Google's LaMDA in PyTorch.
LLaMA
- LLaMA — LLM created by Meta. The model was inadvertently leaked a few days after its announcement. Needless to say, this has been a massive boon for the LLM community. Many optimizations and derivatives of LLaMA have been made, and there are now several Pygmalion models based off of LLaMA (see the Pygmalion Models section above).
- Llama 2 - An improved version of the original LLaMA model, and can now be used for commercial purposes.
- List of LLaMA models
- Subreddit
- Simplified LLaMA installation guide
- The Extremely Simplified First-time LLaMA guide
- LLaMA common questions
- Guide to LLaMA roleplay
- C++ implementation
- Rust implementation
- Llama Hub — A library of custom data loaders/readers for LLMs.
Keep your Smile. Keep on Smiling.