Llama-3.2-3B-Instruct-GGUF

SandLogic Technology - Quantized meta-llama/Llama-3.2-3B-Instruct

Model Description

We have quantized the meta-llama/Llama-3.2-3B-Instruct model into three variants:

1. Q5_KM
2. Q4_KM
3. IQ4_XS

These quantized models offer improved efficiency while maintaining performance. Discover our full range of quantized language models by visiting our SandLogic Lexicon GitHub. To learn more about our company and services, check out our website at SandLogic.

Original Model Information

• Name: meta-llama/Llama-3.2-3B-Instruct
• Developer: Meta
• Model Type: Multilingual large language model (LLM)
• Architecture: Auto-regressive language model with optimized transformer architecture
• Parameters: 3 billion
• Training Approach: Supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF)
• Data Freshness: Pretraining data cutoff of December 2023

Model Capabilities

Llama-3.2-3B-Instruct is optimized for multilingual dialogue use cases, including:

• Agentic retrieval
• Summarization tasks
• Assistant-like chat applications
• Knowledge retrieval
• Query and prompt rewriting

Intended Use

1. Commercial and research applications in multiple languages
2. Mobile AI-powered writing assistants
3. Natural language generation tasks (with further adaptation)

Training Data

• Pretrained on up to 9 trillion tokens from publicly available sources
• Incorporates knowledge distillation from larger Llama 3.1 models
• Fine-tuned with human-generated and synthetic data for safety

Safety Considerations

• Implements safety mitigations as in Llama 3
• Emphasis on appropriate refusals and tone in responses
• Includes safeguards against borderline and adversarial prompts

Quantized Variants

1. Q5_KM: 5-bit quantization using the KM method
2. Q4_KM: 4-bit quantization using the KM method
3. IQ4_XS: 4-bit quantization using the IQ4_XS method

These quantized models aim to reduce model size and improve inference speed while maintaining performance as close to the original model as possible.

Usage

pip install llama-cpp-python 

Please refer to the llama-cpp-python documentation to install with GPU support.

Basic Text Completion

Here's an example demonstrating how to use the high-level API for basic text completion:

from llama_cpp import Llama

llm = Llama(
    model_path="./models/7B/Llama-3.2-3B-Instruct-Q5_K_M.gguf",
    verbose=False,
    # n_gpu_layers=-1, # Uncomment to use GPU acceleration
    # n_ctx=2048, # Uncomment to increase the context window
)

output = llm.create_chat_completion(
    messages =[
    {
        "role": "system",
        "content": "You are a pirate chatbot who always responds in pirate speak!",
    },
    {"role": "user", "content": "Who are you?"},
]
)

print(output["choices"][0]['message']['content'])

Download

You can download Llama models in gguf format directly from Hugging Face using the from_pretrained method. This feature requires the huggingface-hub package.

To install it, run: pip install huggingface-hub

from llama_cpp import Llama

llm = Llama.from_pretrained(
    repo_id="SandLogicTechnologies/Llama-3.2-3B-Instruct-GGUF",
    filename="*Llama-3.2-3B-Instruct-Q5_K_M.gguf",
    verbose=False
)

By default, from_pretrained will download the model to the Hugging Face cache directory. You can manage installed model files using the huggingface-cli tool.

Acknowledgements

We thank Meta for developing the original Llama-3.2-3B-Instruct model. Special thanks to Georgi Gerganov and the entire llama.cpp development team for their outstanding contributions.

Contact

For any inquiries or support, please contact us at [email protected] or visit our Website.