Instructions to use EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- llama-cpp-python
How to use EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF with llama-cpp-python:
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF", filename="Magidonia-24B-v4.3-Q5_0.gguf", )
output = llm( "Once upon a time,", max_tokens=512, echo=True ) print(output)
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF with llama.cpp:
Install from brew
brew install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0 # Run inference directly in the terminal: llama-cli -hf EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0 # Run inference directly in the terminal: llama-cli -hf EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0 # Run inference directly in the terminal: ./llama-cli -hf EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0 # Run inference directly in the terminal: ./build/bin/llama-cli -hf EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0
Use Docker
docker model run hf.co/EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0
- LM Studio
- Jan
- Ollama
How to use EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF with Ollama:
ollama run hf.co/EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0
- Unsloth Studio
How to use EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF to start chatting
- Atomic Chat new
- Docker Model Runner
How to use EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF with Docker Model Runner:
docker model run hf.co/EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0
- Lemonade
How to use EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull EloyOn/Magidonia-24B-v4.3-Q5_0-GGUF:Q5_0
Run and chat with the model
lemonade run user.Magidonia-24B-v4.3-Q5_0-GGUF-Q5_0
List all available models
lemonade list
Partial OG model card
Magidonia 24B v4.3 💿
image
What's New?
Much better roleplay
Might refuse/be more positive
But much better roleplay
Usage
Mistral v7 Tekken
[THINK] [/THINK] might work
works
No thinking also works
Description
The model often surprises me with how good some of the generated answers are. The overall level is also good.
i'm lurking here for weeks, testing out the builds in RP, this is by far the best one. It is pretty smart and aware of it's surroundings and holds the character card really well.
i tried many of your magistral finetunes and for sure this feels the best so nice job^^
I've completed the analysis of v4zk's response, and it's excellent, with virtually no physics errors or contradictory intentions. This is significantly better than previous versions of Cydonia, which always had obvious trade-offs: either too erotic and lacking in physics, or a balance between physics and eroticism at the expense of character development, or simply inconsistency, performing terribly in certain test cases. v4zk's response is virtually error-free and novel, maintaining broad stability while allowing characters to express distinct inner intentions. I believe this is the best version of Cydonia to date. Simply put, it's good enough to approach the level of Skyfall—making up for most of the gains from parameter scaling.
It's a versatile tune, really adheres to system prompts and thinking prefill. OOC, Author's note and all that work flawlessly. Love the thinking on this one! It's good at reading between the lines when given good prompting. It will consider details from the cards in unique ways. Accepts a wide range of sampling parameters and is still stable (comparatively, it may be too creative at high temperatures). Swipes are minimal compared to other similarly-sized thinking models. Good for RP/CYOA/Dungeon/Chat/Companion/Author!
You can also try its Mistral Small 3.2 counterpart: Cydonia 24B v4.3 (The difference is mostly writing style, and most prefer the Magistral tune.)