---
tags:
- sentinel-manifold
- quantization
- edge-deployment
- int8
- pytorch
license: mit
---

# 🦴 Sentinel Tiny Text — Sentinel-INT8

**Production-ready quantized model for edge deployment.**

This is the Sentinel-INT8 quantized variant of [Sentinel Tiny Text](https://huggingface.co/5dimension/sentinel-tiny-text).

---

## 📊 Quantization Spec (SDM-v1 §4)

| Property | Value |
|----------|-------|
| **Format** | Sentinel-INT8 |
| **Zero-Point (Z)** | C₁ = -0.007994021806 (attracting fixed point) |
| **Scale (S)** | max\|w\| · (1/e) |
| **Compression** | 4× |
| **Size** | ~13 MB |
| **Parent model** | [5dimension/sentinel-tiny-text](https://huggingface.co/5dimension/sentinel-tiny-text) |

---

## 🧬 Mathematical Foundation

The Sentinel quantization uses the attracting fixed point C₁ as the zero-point, minimizing rounding noise by clustering weights around the dynamical attractor.

```
Quantization: q = round((w - C₁) / S)
Dequantization: w = q · S + C₁
Where S = max|w| · (1/e)
```

---

## 🚀 Deployment Targets

| Platform | Compatible |
|----------|-----------|
| Mobile (iOS/Android) | ✅ |
| IoT devices | ✅ |
| Microcontrollers (ARM Cortex-M) | ✅ |
| Browser (ONNX/WebGL) | ✅ |
| Edge TPU | ⚠️ (needs conversion) |

---

## 📦 Download

```python
from huggingface_hub import hf_hub_download

# Download quantized weights
weights = hf_hub_download(repo_id="5dimension/sentinel-tiny-text-int8", filename="model_int8.pt")
```

---

## 🔗 Links

- **Parent model**: [5dimension/sentinel-tiny-text](https://huggingface.co/5dimension/sentinel-tiny-text)
- **Main repo**: [sentinel-manifold-discoveries](https://huggingface.co/5dimension/sentinel-manifold-discoveries)
- **Sentinel Hub**: [Interactive demo](https://huggingface.co/spaces/5dimension/sentinel-hub)

---

**License:** MIT | **Edge-ready. Theorem-backed.** 🦴