Metadata-Version: 2.4
Name: cpn-engine
Version: 2.5.0
Summary: Contextual Precision Network (CPN) Engine — Dynamic Mixed-Precision LLM Inference Engine
Home-page: https://github.com/OmerNL34/cpn-engine
Author: OmerNL34
Author-email: omerwindnl@gmail.com
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.10
Description-Content-Type: text/markdown
Requires-Dist: torch>=2.0.0
Requires-Dist: transformers>=4.36.0
Requires-Dist: accelerate>=0.25.0
Requires-Dist: numpy
Dynamic: author
Dynamic: author-email
Dynamic: classifier
Dynamic: description
Dynamic: description-content-type
Dynamic: home-page
Dynamic: requires-dist
Dynamic: requires-python
Dynamic: summary

# 🚀 CPN Engine (v2.5) — Contextual Precision Network

[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
[![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
[![PyTorch 2.0+](https://img.shields.io/badge/PyTorch-2.0+-EE4C2C.svg)](https://pytorch.org/)
[![VRAM Savings](https://img.shields.io/badge/VRAM%20Savings-Up%20to%2053.1%25-green.svg)](#benchmark--performance)

**CPN Engine** is an enterprise-grade inference engine that dynamically optimizes LLMs using Contextual Mixed-Precision Routing and Adaptive Entropy Thresholding.

## 📊 Benchmark & Performance (RTX 4070 Ti - 12GB VRAM)

| Model Architecture | Baseline Precision | CPN Mode | Baseline VRAM | CPN Peak VRAM | VRAM Savings |
| :--- | :---: | :---: | :---: | :---: | :---: |
| **Qwen2.5-3B** | FP16 | CPN Hybrid (16/4-bit) | 6.20 GB | **3.95 GB** | **%36.2** |
| **DeepSeek-R1-Distill-1.5B** | FP16 | CPN Hybrid (16/4-bit) | 3.31 GB | **2.84 GB** | **%14.3** |
| **DeepSeek-R1-Distill-7B** | FP16 | CPN Hybrid (16/4-bit) | 14.12 GB (OOM) | **6.63 GB** | **%53.1** |

## 💻 Quick Start




