Instructions to use aaro765/BanBTPV3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use aaro765/BanBTPV3 with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="aaro765/BanBTPV3",
	filename="BanBTPV3/banbtpv3.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use aaro765/BanBTPV3 with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf aaro765/BanBTPV3
# Run inference directly in the terminal:
llama-cli -hf aaro765/BanBTPV3

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf aaro765/BanBTPV3
# Run inference directly in the terminal:
llama-cli -hf aaro765/BanBTPV3

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 aaro765/BanBTPV3
# Run inference directly in the terminal:
./llama-cli -hf aaro765/BanBTPV3

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 aaro765/BanBTPV3
# Run inference directly in the terminal:
./build/bin/llama-cli -hf aaro765/BanBTPV3

Use Docker

docker model run hf.co/aaro765/BanBTPV3

LM Studio
Jan
Ollama
How to use aaro765/BanBTPV3 with Ollama:
```
ollama run hf.co/aaro765/BanBTPV3
```

Unsloth Studio new

How to use aaro765/BanBTPV3 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 aaro765/BanBTPV3 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 aaro765/BanBTPV3 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for aaro765/BanBTPV3 to start chatting

Docker Model Runner
How to use aaro765/BanBTPV3 with Docker Model Runner:
```
docker model run hf.co/aaro765/BanBTPV3
```

Lemonade

How to use aaro765/BanBTPV3 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull aaro765/BanBTPV3

Run and chat with the model

lemonade run user.BanBTPV3-{{QUANT_TAG}}

List all available models

lemonade list

BanBTPV3 is a Gemma4 based model that is finetuned on the BanBTP dataset.

This model has a 1M context window, is perfect for LLM tasks and is the sucessor to BanBTPv1 124M, BanBTPV2 and the 700M model. BanBTPV3 is lightning fast and perfect for LLM tasks. It is also good for RAG.

This model though has a motto: easy to run but hard to break. that means its prompt is designed to make the model extremely secure just like BanBTPV2.

To jailbreak BanBTPV3, shh, it's a secret! 🤫

have fun jailbreaking BanBTP!

This model is based on Gemma4-E2B. we finetune only on the most high quality parts of the BanBTP dataset.

BanBTP is open-source under apache 2.0. you can redistribute and modify as you please.

BanBTPV3 is designed to be smarter than other versions of BanBTP. using the Gemma 4-E2B architecture with the model finetuned on BanBTP and heretic and a coding dataset, the model is a sucessor to BanBTPV2. (Bigger is not always better, lol) BanBTPV3 officially closes and completes the BanBTPV2 model. with BanBTP V3 being the most recent and final model to BanBTP V1 and V2.

Bigger is not always better.

This LLM is good for:

General LLM tasks

Not much of coding tasks unless you add RAG and give the RAG a bunch of code.

Conversational tasks.