Instructions to use fevohh/RayExtract-3B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use fevohh/RayExtract-3B with Transformers:

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("fevohh/RayExtract-3B", dtype="auto")

llama-cpp-python

How to use fevohh/RayExtract-3B with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="fevohh/RayExtract-3B",
	filename="unsloth.Q4_K_M.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 fevohh/RayExtract-3B with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf fevohh/RayExtract-3B:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf fevohh/RayExtract-3B:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf fevohh/RayExtract-3B:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf fevohh/RayExtract-3B: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 fevohh/RayExtract-3B:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf fevohh/RayExtract-3B: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 fevohh/RayExtract-3B:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf fevohh/RayExtract-3B:Q4_K_M

Use Docker

docker model run hf.co/fevohh/RayExtract-3B:Q4_K_M

LM Studio
Jan
Ollama
How to use fevohh/RayExtract-3B with Ollama:
```
ollama run hf.co/fevohh/RayExtract-3B:Q4_K_M
```

Unsloth Studio new

How to use fevohh/RayExtract-3B 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 fevohh/RayExtract-3B 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 fevohh/RayExtract-3B to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for fevohh/RayExtract-3B to start chatting

Pi new

How to use fevohh/RayExtract-3B with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf fevohh/RayExtract-3B:Q4_K_M

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": "RayExtract-3B"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use fevohh/RayExtract-3B with Docker Model Runner:
```
docker model run hf.co/fevohh/RayExtract-3B:Q4_K_M
```

Lemonade

How to use fevohh/RayExtract-3B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull fevohh/RayExtract-3B:Q4_K_M

Run and chat with the model

lemonade run user.RayExtract-3B-Q4_K_M

List all available models

lemonade list

Remarks

The model is capable of extracting json output accurately i.e. listing item price as an integer if price is mentioned or "na" if there is no price, it is also able to handle unclean data inputs outside of the dataset. However, its context outputs are concise and lack further reasoning and does not fully follow the thinking steps provided from the dataset, such as not implementing item currency from a given price, which led to a lower accuracy extract output but still maintains high quality json format output.