Instructions to use MalithaBandara/EleGuard with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- llama-cpp-python
How to use MalithaBandara/EleGuard with llama-cpp-python:
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="MalithaBandara/EleGuard", filename="gemma-4-e2b-it.F16-mmproj.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 MalithaBandara/EleGuard with llama.cpp:
Install from brew
brew install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf MalithaBandara/EleGuard:F16 # Run inference directly in the terminal: llama-cli -hf MalithaBandara/EleGuard:F16
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf MalithaBandara/EleGuard:F16 # Run inference directly in the terminal: llama-cli -hf MalithaBandara/EleGuard: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 MalithaBandara/EleGuard:F16 # Run inference directly in the terminal: ./llama-cli -hf MalithaBandara/EleGuard: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 MalithaBandara/EleGuard:F16 # Run inference directly in the terminal: ./build/bin/llama-cli -hf MalithaBandara/EleGuard:F16
Use Docker
docker model run hf.co/MalithaBandara/EleGuard:F16
- LM Studio
- Jan
- Ollama
How to use MalithaBandara/EleGuard with Ollama:
ollama run hf.co/MalithaBandara/EleGuard:F16
- Unsloth Studio new
How to use MalithaBandara/EleGuard 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 MalithaBandara/EleGuard 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 MalithaBandara/EleGuard to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for MalithaBandara/EleGuard to start chatting
- Docker Model Runner
How to use MalithaBandara/EleGuard with Docker Model Runner:
docker model run hf.co/MalithaBandara/EleGuard:F16
- Lemonade
How to use MalithaBandara/EleGuard with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull MalithaBandara/EleGuard:F16
Run and chat with the model
lemonade run user.EleGuard-F16
List all available models
lemonade list
Update README.md
Browse files
README.md
CHANGED
|
@@ -4,4 +4,49 @@ language:
|
|
| 4 |
- en
|
| 5 |
base_model:
|
| 6 |
- google/gemma-4-E2B-it
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
---
|
|
|
|
| 4 |
- en
|
| 5 |
base_model:
|
| 6 |
- google/gemma-4-E2B-it
|
| 7 |
+
tags:
|
| 8 |
+
- gemma
|
| 9 |
+
- gguf
|
| 10 |
+
- multimodal
|
| 11 |
+
- vision
|
| 12 |
+
- wildlife-monitoring
|
| 13 |
+
- quantized
|
| 14 |
+
- audio
|
| 15 |
+
- text
|
| 16 |
+
---
|
| 17 |
+
# EleGuard: Multimodal Elephant Detection & Reasoning
|
| 18 |
+
|
| 19 |
+
**EleGuard** is a specialized, multimodal Vision-Language Model (VLM) developed for the **24/7 monitoring of elephant activity** in natural habitats. By leveraging infrared (IR) imagery and bioacoustic signals, EleGuard provides a robust solution for human-elephant conflict mitigation and wildlife conservation.
|
| 20 |
+
|
| 21 |
+
## Model Summary
|
| 22 |
+
* **Distinct Project Name:** EleGuard
|
| 23 |
+
* **Base Architecture:** This model is a variant based on **Gemma 4 E2B**.
|
| 24 |
+
* **Modality:** Multimodal (Vision + Acoustic via Spectrograms).
|
| 25 |
+
* **Format:** GGUF (Optimized for edge deployment).
|
| 26 |
+
* **Training Method:** Knowledge Distillation from Gemini 3.1 Flash.
|
| 27 |
+
|
| 28 |
+
## Technical Innovation: Reasoning Distillation
|
| 29 |
+
The core breakthrough of EleGuard is the shift from simple classification to **expert reasoning**. Instead of training only on labels, the model was fine-tuned on "thought blocks" generated by a Teacher model (Gemini 3.1 Flash).
|
| 30 |
+
|
| 31 |
+
For every image or audio sample, the model is trained to explain its reasoning—such as identifying thermal signatures in thick brush or frequency patterns in a rumble—before outputting a final status:
|
| 32 |
+
* **🚨 ALERT:** Elephant presence confirmed.
|
| 33 |
+
* **✅ SAFE:** No threat detected.
|
| 34 |
+
|
| 35 |
+
## Dataset Details
|
| 36 |
+
The model was trained on a curated dataset of **2,600 samples** organized into:
|
| 37 |
+
* **Visual Imagery:** High-resolution daytime and **Infrared (IR)** forest captures.
|
| 38 |
+
* **Acoustic Data:** Mel Spectrograms identifying vocalizations like rumbles, roars, and trumpets.
|
| 39 |
+
* **Paired Expert Labels:** Detailed JSON reasoning files for every media asset.
|
| 40 |
+
|
| 41 |
+
## Usage & Deployment
|
| 42 |
+
This repository contains the model weights in **GGUF** format, specifically optimized for edge devices (Raspberry Pi, Jetson Nano, or standard laptops) using tools like `llama.cpp` or `Ollama`.
|
| 43 |
+
|
| 44 |
+
### Required Files:
|
| 45 |
+
1. `EleGuard-gemma-4-e2b-it.GGUF` (Main model weights)
|
| 46 |
+
2. `EleGuard-gemma-4-e2b-it.mmproj.GGUF` (Multimodal vision projector)
|
| 47 |
+
|
| 48 |
+
## Acknowledgments & Trademarks
|
| 49 |
+
* **Gemma is a trademark of Google LLC.**
|
| 50 |
+
* EleGuard is a model trained on a dataset based on Gemma 4 E2B.
|
| 51 |
+
* This project was developed for [The Gemma 4 Good Hackathon] using the Unsloth fine-tuning framework.
|
| 52 |
---
|