Termux-Vision
디바이스 리소스를 활용한 안드로이드 Termux 제로 디펜던시 온디바이스 컴퓨터 비전 & VLM 멀티모달 추론 엔진
Zero-Dependency On-Device Computer Vision & Multimodal VLM Inference Engine Utilizing Device Resources for Android Termux
📌 Architecture & Overview
순수 ARM64 NEON 커널과 4단계 화질 프리셋(fast, optimal, high, original), 디바이스 리소스 최적화 연산을 결합하여 메모리 격리 환경에서 엣지 검출과 VLM 멀티모달 시각 추론을 실현합니다.
Eliminates heavy C++ dependencies by integrating SIMD NEON spatial image transforms with on-device VLM (Qwen2-VL, SmolVLM) multi-tier resolution presets (fast, optimal, high, original) utilizing device resources under strict memory isolation.
🚀 Installation & Quickstart
Python (PyPI)
pip install termux-vision
import termux_vision as tv
# 1. Zero-Dependency Spatial Filtering (0.01s ultra-fast)
img = tv.io.load_image("photo.jpg")
edges = tv.cv.canny(tv.transforms.to_grayscale(img), 40, 120)
# 2. On-Device VLM Multimodal Inference with 4-Tier Quality Presets
with tv.vlm.load("qwen2-vl-2b-q4", quality="optimal") as engine:
res = engine.describe("photo.jpg", prompt="Describe this scene in detail.", quality="optimal")
print(f"Generated ({res.metrics.tokens_per_second:.1f} t/s): {res.text}")
Node.js / TypeScript (npm)
npm install termux-vision
import tv from 'termux-vision';
// 1. Diagnostics & Hardware Probe
const doc = tv.doctor(true);
console.log(`Vulkan GPU: ${doc.vulkan.status} | RAM: ${doc.hardware.availableRamMb} MB`);
// 2. Multimodal VLM Inference with Quality Preset
const engine = await tv.load({ modelId: 'qwen2-vl-2b-q4', contextLimit: 4096 });
const result = await engine.describe('photo.jpg', { quality: 'optimal', maxTokens: 300 });
console.log(`[${result.metrics.backend.toUpperCase()}] ${result.text}`);
engine.close();
📖 Official Documentation & Benchmarks
- Official Architecture & API Reference
- Ecosystem Metrics & Registry Stats
- AMEVA Open-Source Foundation Portal
📄 License
Licensed under the Apache-2.0 License. Copyright (c) 2026 Eunho Kim (@uno-km).
1 maintainer
Metadata
1 maintainer
Release files for termux-vision 1.1.4
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| termux_vision-1.1.4.tar.gz | 84.3 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| termux_vision-1.1.4-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 172.4 kB
Release files / termux_vision-1.1.4.tar.gz
| Download URL | termux_vision-1.1.4.tar.gz |
|---|---|
| Size | 84.3 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
e2fd44f7cfc09b9a4033f72d4af8a65fdb4f71f1e950500a56e143441b4804b7
|
|
BLAKE2b-256 checksum How to use checksums |
a9abf1166b2d5e6124c02a45bd165ee2061af3f1240416c45824aff9a0604a36
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.0
|
Release files / termux_vision-1.1.4-py3-none-any.whl
| Download URL | termux_vision-1.1.4-py3-none-any.whl |
|---|---|
| Size | 88.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
83a1ad7fa0fb4c91d00f0291b58cabf575a56a3fb2efd1d3367ecf2511446929
|
|
BLAKE2b-256 checksum How to use checksums |
a073fbf50f33f957a21211f41a2a588e2892b4807a6f32187bb7991e939c40e9
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.0
|
