File size: 1,369 Bytes
2d1c2b5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8e72432
2d1c2b5
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
---
# Model card metadata following Hugging Face specification: 
# https://github.com/huggingface/hub-docs/blob/main/modelcard.md?plain=1
# Documentation: https://huggingface.co/docs/hub/model-cards
license: mit
tags:
  - llama-cpp-python
  - cuda
  - nvidia
  - blackwell
  - windows
  - prebuilt-wheels
  - python
  - machine-learning
  - large-language-models
  - gpu-acceleration
---

# llama-cpp-python 0.3.9 Prebuilt Wheel with CUDA Support for Windows

This repository provides a prebuilt Python wheel for **llama-cpp-python** (version 0.3.9) with NVIDIA CUDA support, for Windows 10/11 (x64) systems. This wheel enables GPU-accelerated inference for large language models (LLMs) using the `llama.cpp` library, simplifying setup by eliminating the need to compile from source. The wheel is compatible with Python 3.10 and supports NVIDIA GPUs, including the latest Blackwell architecture.

## Available Wheel
- `llama_cpp_python-0.3.9-cp310-cp310-win_amd64.whl` (Python 3.10, CUDA 12.8)

## Compatibility
The prebuilt wheels are designed for NVIDIA Blackwell GPUs but have been tested and confirmed compatible with previous-generation NVIDIA GPUs, including:
- NVIDIA RTX 5090
- NVIDIA RTX 3090

## Installation
To install the wheel, use the following command in your Python 3.10 environment:

```bash
pip install llama_cpp_python-0.3.9-cp310-cp310-win_amd64.whl