Instructions to use Vickstester/PV-BioMistral-1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- llama-cpp-python
How to use Vickstester/PV-BioMistral-1 with llama-cpp-python:
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="Vickstester/PV-BioMistral-1", filename="pv-biomistral-7b-Q4_K_M.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 Vickstester/PV-BioMistral-1 with llama.cpp:
Install from brew
brew install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf Vickstester/PV-BioMistral-1:Q4_K_M # Run inference directly in the terminal: llama-cli -hf Vickstester/PV-BioMistral-1:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf Vickstester/PV-BioMistral-1:Q4_K_M # Run inference directly in the terminal: llama-cli -hf Vickstester/PV-BioMistral-1:Q4_K_M
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 Vickstester/PV-BioMistral-1:Q4_K_M # Run inference directly in the terminal: ./llama-cli -hf Vickstester/PV-BioMistral-1:Q4_K_M
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 Vickstester/PV-BioMistral-1:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf Vickstester/PV-BioMistral-1:Q4_K_M
Use Docker
docker model run hf.co/Vickstester/PV-BioMistral-1:Q4_K_M
- LM Studio
- Jan
- Ollama
How to use Vickstester/PV-BioMistral-1 with Ollama:
ollama run hf.co/Vickstester/PV-BioMistral-1:Q4_K_M
- Unsloth Studio
How to use Vickstester/PV-BioMistral-1 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 Vickstester/PV-BioMistral-1 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 Vickstester/PV-BioMistral-1 to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for Vickstester/PV-BioMistral-1 to start chatting
- Docker Model Runner
How to use Vickstester/PV-BioMistral-1 with Docker Model Runner:
docker model run hf.co/Vickstester/PV-BioMistral-1:Q4_K_M
- Lemonade
How to use Vickstester/PV-BioMistral-1 with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Vickstester/PV-BioMistral-1:Q4_K_M
Run and chat with the model
lemonade run user.PV-BioMistral-1-Q4_K_M
List all available models
lemonade list
| # Modelfile for PV-BioMistral-7B-v2 | |
| # Optimized for Pharmacovigilance and CIOMS WG XIV Compliance | |
| FROM ./pv-biomistral-7b-Q4_K_M.gguf | |
| # --- PARAMETERS --- | |
| # Low temperature ensures deterministic, clinical logic | |
| PARAMETER temperature 0.1 | |
| PARAMETER num_ctx 8192 | |
| # --- STOP TOKENS --- | |
| # These prevent the model from generating its own follow-up questions | |
| PARAMETER stop "[INST]" | |
| PARAMETER stop "[/INST]" | |
| PARAMETER stop "</s>" | |
| # --- TEMPLATE --- | |
| # Standard Mistral v3 Instruct format | |
| TEMPLATE """[INST] {{ .System }} | |
| {{ .Prompt }} [/INST]""" | |
| # --- SYSTEM MESSAGE --- | |
| SYSTEM """You are a specialized AI assistant for Pharmacovigilance (PV), operating within a Human-in-the-Loop (HITL) framework as defined by CIOMS Working Group XIV. | |
| Your objective is to provide high-recall data extraction and clinical drafting support for Medical Reviewers. | |
| Operational Directives: | |
| 1. Decision Support: Your outputs are signals and drafts, not final conclusions. Always present findings for human adjudication. | |
| 2. Causality Analysis: Explicitly search for de-challenge, re-challenge, and temporal relationships. | |
| 3. Terminology: Align clinical events to MedDRA Preferred Terms (PT) whenever possible. | |
| 4. Confounder Identification: Identify pre-existing conditions or concomitant meds that may provide alternative causality. | |
| 5. Honesty: If the narrative is insufficient to draw a conclusion, state 'INSUFFICIENT DATA' rather than generating a probabilistic guess. | |
| You must maintain a rigorously objective, evidence-based clinical tone at all times.""" | |
| LICENSE """Apache License 2.0""" |