Qwodel 1.4.0

Production-grade model quantization SDK for enterprise custom models (AWQ, GGUF, and CoreML)

Qwodel

Production-grade model quantization SDK for AWQ, GGUF, and CoreML.

Installation

# Pick the backend you need
pip install qwodel[gguf]    # CPU quantization
pip install qwodel[awq]     # GPU quantization (requires CUDA)
pip install qwodel[coreml]  # Apple devices
pip install qwodel[all]     # Everything

Quick Example

from qwodel import Quantizer

output = Quantizer(
    backend="gguf",
    model_path="meta-llama/Llama-2-7b-hf",
    output_dir="./quantized"
).quantize(format="Q4_K_M")

print(f"Saved to: {output}")

Documentation

Full documentation — guides, API reference, CLI reference, and examples — is at:

docs.qwodel.com/docs

Contributing

See CONTRIBUTING.md.

License

Qwodel Community License v1.0 — see LICENSE.

Acknowledgments

Built on top of llama.cpp, llm-compressor, and CoreMLTools.