EFEai Gemma 2 (2B) On-Device Mobile Engine
This repository hosts the mobile-optimized, 4-bit quantized Google Gemma 2 (2B) model weights for the Kanal Analiz & EFEai Android application by Yağız Efe Gülenoğlu.
- Architecture: Gemma 2 (2B) Instruct
- Quantization: INT4 (Optimized for Mobile NPU/GPU/CPU)
- Target Runtime: Google MediaPipe Tasks GenAI / LiteRT
- Target Device: Android 8GB+ RAM devices
- Offline Capable: Yes (100% On-Device, zero cloud API calls)