https://github.com/0xhilsa/tenop
A lightweight & minimalist tensor computation library with CUDA backend
https://github.com/0xhilsa/tenop
bash c cuda python3 tensor
Last synced: 2 months ago
JSON representation
A lightweight & minimalist tensor computation library with CUDA backend
- Host: GitHub
- URL: https://github.com/0xhilsa/tenop
- Owner: 0xhilSa
- License: mit
- Created: 2025-04-21T20:16:03.000Z (about 1 year ago)
- Default Branch: master
- Last Pushed: 2025-06-09T20:44:24.000Z (about 1 year ago)
- Last Synced: 2025-06-09T21:29:55.574Z (about 1 year ago)
- Topics: bash, c, cuda, python3, tensor
- Language: Python
- Homepage:
- Size: 349 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
TenOp is a lightweight and minimalist tensor computation library with CUDA acceleration, designed for high-performance numerical computing.
---
## 🚀 Features
- Simple tensor operations with a clean API
- Optimized backend with CPU and CUDA support
- Written in Python with compiled C/CUDA extensions
## 📦 Installation
```bash
pip install tenop
```
## Or install from source
```bash
git clone https://github.com/0xhilSa/tenop.git
cd tenop
pip install .
```
## Usage
```python
from tenop import Tensor
x = Tensor([1,2,3,4,5], device="cpu:0")
print(x)
print(x.numpy())
print(x.device, x.device.name)
y = x.cuda()
print(y)
print(y.numpy())
print(y.device, y.device.name)
```
See [./tests/testing.ipynb](./tests/testing.ipynb) for more examples
---
## LICENSE
[MIT](./LICENSE)