---
language: en
license: apache-2.0
tags:
  - aqea
  - compression
  - embeddings
  - similarity-search
  - vector-database
datasets:
  - mteb/stsbenchmark-sts
base_model: sentence-transformers/all-MiniLM-L6-v2
---

# AQEA: aqea-all-MiniLM-L6-v2-29x

MiniLM text embeddings compressed 29x while preserving 97.1% similarity ranking

## 📊 Performance

| Metric | Value |
|--------|-------|
| **Compression Ratio** | 29.5x |
| **Spearman ρ** | 97.1% |
| **Source Dimension** | 384D |
| **Compressed Dimension** | 13D |
| **Storage Savings** | 96.6% |

## 🚀 Usage

```python
from aqea import AQEACompressor

# Load pre-trained compressor
compressor = AQEACompressor.from_pretrained("nextxag/aqea-all-MiniLM-L6-v2-29x")

# Compress embeddings
embeddings = model.encode(texts)  # 384D
compressed = compressor.compress(embeddings)  # 13D

# Decompress for retrieval
reconstructed = compressor.decompress(compressed)  # 384D
```

## 📁 Files

- `weights.aqwt` - Binary weights (AQEA native format)
- `config.json` - Model configuration

## 🔬 How It Works

AQEA (Adaptive Quantized Embedding Architecture) uses learned linear projections 
with Pre-Quantify rotation to compress embeddings while maximally preserving 
pairwise similarity rankings (measured by Spearman correlation).

## 📚 Citation

```bibtex
@software{aqea2024,
  title = {AQEA: Adaptive Quantized Embedding Architecture},
  author = {AQEA Team},
  year = {2024},
  url = {https://huggingface.co/nextxag}
}
```

## 📄 License

Apache 2.0