Instructions to use mhylle/ministral-3b-mist-9liner-bilingual-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use mhylle/ministral-3b-mist-9liner-bilingual-gguf 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 mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16
# Run inference directly in the terminal:
llama cli -hf mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16
# Run inference directly in the terminal:
llama cli -hf mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

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 mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16
# Run inference directly in the terminal:
./llama-cli -hf mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

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 mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

Use Docker

docker model run hf.co/mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

LM Studio
Jan

vLLM

How to use mhylle/ministral-3b-mist-9liner-bilingual-gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mhylle/ministral-3b-mist-9liner-bilingual-gguf"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mhylle/ministral-3b-mist-9liner-bilingual-gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

Ollama
How to use mhylle/ministral-3b-mist-9liner-bilingual-gguf with Ollama:
```
ollama run hf.co/mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16
```
Unsloth Desktop

How to use mhylle/ministral-3b-mist-9liner-bilingual-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use mhylle/ministral-3b-mist-9liner-bilingual-gguf with Docker Model Runner:
```
docker model run hf.co/mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16
```

Lemonade

How to use mhylle/ministral-3b-mist-9liner-bilingual-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

Run and chat with the model

lemonade run user.ministral-3b-mist-9liner-bilingual-gguf-F16

List all available models

lemonade list

Hermes Agent

How to use mhylle/ministral-3b-mist-9liner-bilingual-gguf with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

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 mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use mhylle/ministral-3b-mist-9liner-bilingual-gguf with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16

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 "mhylle/ministral-3b-mist-9liner-bilingual-gguf:F16" \
  --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"

mhylle commited on Jan 19

Commit

04796aa

verified ·

1 Parent(s): 06af3be

Add model card

Browse files

Files changed (1) hide show

README.md +86 -0

README.md ADDED Viewed

	@@ -0,0 +1,86 @@

+---
+license: apache-2.0
+language:
+  - en
+  - de
+library_name: gguf
+base_model: mistralai/Ministral-3-3B-Instruct-2512-BF16
+tags:
+  - medical
+  - dictation
+  - structured-extraction
+  - mist-9liner
+  - on-device
+  - gguf
+  - ministral
+pipeline_tag: text-generation
+datasets:
+  - custom
+---
+# Ministral 3B - MIST 9-Liner Bilingual (GGUF)
+This is a **GGUF format** export of a fine-tuned Ministral 3B model for medical dictation extraction.
+## Model Description
+- **Base Model**: mistralai/Ministral-3-3B-Instruct-2512-BF16
+- **Training Method**: LoRA (merged with base model)
+- **Languages**: English + German (bilingual)
+- **Task**: Extract structured MIST 9-liner data from medical dictations
+- **Format**: GGUF F16 (for use with llama.cpp)
+## Performance Metrics
+| Metric | Score |
+|--------|-------|
+| JSON Valid | 100% |
+| Schema Compliant | 98% |
+| Avg Field Accuracy | 69.5% |
+## Training Details
+- **Epochs**: 2
+- **Batch Size**: 2 (effective 16 with gradient accumulation)
+- **Learning Rate**: 1.5e-4
+- **Max Sequence Length**: 4096
+- **LoRA Rank**: 16
+- **LoRA Alpha**: 32
+## Model Comparison
+| Model | Size | JSON Valid | Schema | Field Accuracy |
+|-------|------|------------|--------|----------------|
+| **Ministral 3B** | 6.4GB | 100% | 98% | 69.5% |
+| Gemma3 270M | 536MB | 99% | 96% | 68.7% |
+## Intended Use
+This model is designed for on-device deployment to extract structured MIST 9-liner information from medical dictation text. The MIST 9-liner format includes:
+1. **M**echanism of injury
+2. **I**njury pattern
+3. **S**igns and symptoms
+4. **T**reatment given
+5. Vital signs (5 fields)
+## Usage with llama.cpp
+```bash
+# Download the GGUF file
+huggingface-cli download mhylle/ministral-3b-mist-9liner-bilingual-gguf ministral-3b-mist-bilingual-f16.gguf
+# Run inference
+./llama-cli -m ministral-3b-mist-bilingual-f16.gguf -p "<prompt>"
+```
+## Limitations
+- Optimized for MIST 9-liner format specifically
+- Best performance with clear, structured dictation input
+- Trained on synthetic/curated medical dictation data
+- Larger model size (6.4GB) may require more memory than smaller alternatives
+## Model Card Contact
+For questions or issues, please open a GitHub issue on the project repository.