Instructions to use baten-ai/LISAN-M7B-4L with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use baten-ai/LISAN-M7B-4L with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf baten-ai/LISAN-M7B-4L # Run inference directly in the terminal: llama cli -hf baten-ai/LISAN-M7B-4L
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf baten-ai/LISAN-M7B-4L # Run inference directly in the terminal: llama cli -hf baten-ai/LISAN-M7B-4L
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 baten-ai/LISAN-M7B-4L # Run inference directly in the terminal: ./llama-cli -hf baten-ai/LISAN-M7B-4L
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 baten-ai/LISAN-M7B-4L # Run inference directly in the terminal: ./build/bin/llama-cli -hf baten-ai/LISAN-M7B-4L
Use Docker
docker model run hf.co/baten-ai/LISAN-M7B-4L
- LM Studio
- Jan
- vLLM
How to use baten-ai/LISAN-M7B-4L with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "baten-ai/LISAN-M7B-4L" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "baten-ai/LISAN-M7B-4L", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/baten-ai/LISAN-M7B-4L
- Ollama
How to use baten-ai/LISAN-M7B-4L with Ollama:
ollama run hf.co/baten-ai/LISAN-M7B-4L
- Unsloth Studio
How to use baten-ai/LISAN-M7B-4L 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 baten-ai/LISAN-M7B-4L 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 baten-ai/LISAN-M7B-4L to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for baten-ai/LISAN-M7B-4L to start chatting
- Pi
How to use baten-ai/LISAN-M7B-4L with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf baten-ai/LISAN-M7B-4L
Configure the model in Pi
# Install Pi: npm install -g @mariozechner/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "baten-ai/LISAN-M7B-4L" } ] } } }Run Pi
# Start Pi in your project directory: pi
- OpenClaw new
How to use baten-ai/LISAN-M7B-4L with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf baten-ai/LISAN-M7B-4L
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "baten-ai/LISAN-M7B-4L" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
- Docker Model Runner
How to use baten-ai/LISAN-M7B-4L with Docker Model Runner:
docker model run hf.co/baten-ai/LISAN-M7B-4L
- Lemonade
How to use baten-ai/LISAN-M7B-4L with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull baten-ai/LISAN-M7B-4L
Run and chat with the model
lemonade run user.LISAN-M7B-4L-{{QUANT_TAG}}List all available models
lemonade list
- Hermes Agent
How to use baten-ai/LISAN-M7B-4L with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf baten-ai/LISAN-M7B-4L
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default baten-ai/LISAN-M7B-4L
Run Hermes
hermes
- Atomic Chat
LISAN-M7B-4L
One question, four minds.
The same model reasons in four registers — it decides, explores, founds, connects — and runs entirely on your machine, offline. (LISAN answers in French.)
The four registers
Ask them the same question and you get four distinct, clear ways of thinking.
| Variant | Register | In a word |
|---|---|---|
| 4La | Decisive | cuts, prioritizes, acts — concise, directive |
| 4Lb | Exploratory | multiple angles, hypotheses, questions, links |
| 4Lc | Foundational | definitions, principles, invariants, rigor |
| 4Ld | Systemic | network, interdependencies, whole-system effects, emergence |
These registers are orthogonal — four independent directions, not four notches of one dial. (The internal construction that generates them is proprietary.)
Example — "A team is starting a complex project. Where to begin?" (answers in French)
- 4La : « On choisit le point d'impact le plus rapide. On tranche : une livraison minimale en 30 jours, puis on avance. »
- 4Lb : « Clarifie d'abord la fin. Quels risques ? quelles dépendances ? Un projet complexe est une bataille d'alignement avant d'être technique. »
- 4Lc : « Premier principe : un socle stable. Trois fondations — vision, dépendances, métrique — avant toute action. »
- 4Ld : « Cartographie les flux de valeur et les interdépendances. Un projet émerge d'un réseau, pas d'une ligne. »
Sovereign and local
- Sovereign — runs 100% locally, no data leaves your machine.
- Model-agnostic — it's the in-house model of the BATEN ecosystem, which governs any other LLM just as well.
See governance in action: chat.baten.ai.
How the roles split
LISAN provides the register — the right way to reason. Traceability and reliability come from the BATEN layer: deterministic routing and a causal seal that attaches verifiable provenance to every response (causalseal.org). The model reasons; the governance proves.
Files
Q4_K_M quantization (GGUF, ~4.4 GB each), base Mistral-7B-Instruct-v0.3.
| File | Register |
|---|---|
lisan-m7b-4l-a-q4km.gguf |
4La — Decisive |
lisan-m7b-4l-b-q4km.gguf |
4Lb — Exploratory |
lisan-m7b-4l-c-q4km.gguf |
4Lc — Foundational |
lisan-m7b-4l-d-q4km.gguf |
4Ld — Systemic |
Usage (Ollama)
Each variant carries its register in its weights. Prompt template: Mistral-Instruct ([INST] … [/INST]).
ollama create lisan-4la -f Modelfile.4la
ollama run lisan-4la "Une équipe démarre un projet complexe. Par où commencer ?"
One Modelfile per variant is provided (Modelfile.4la … Modelfile.4ld).
License & attribution
Apache 2.0. Derived from mistralai/Mistral-7B-Instruct-v0.3 (Apache 2.0) — attribution preserved.
🇫🇷 En français
Une question, quatre esprits. LISAN raisonne en quatre registres — il décide, explore, fonde, relie — tourne 100 % en local et répond en français. C'est le modèle souverain de l'écosystème BATEN : LISAN raisonne, la gouvernance prouve (routage déterministe + sceau causal, provenance vérifiable — causalseal.org). Les quatre registres — 4La décisif · 4Lb exploratoire · 4Lc fondationnel · 4Ld systémique — sont orthogonaux. Détails, exemples et Modelfiles : voir la version anglaise ci-dessus.
- Downloads last month
- 146
We're not able to determine the quantization variants.
Model tree for baten-ai/LISAN-M7B-4L
Base model
mistralai/Mistral-7B-v0.3