TranslateGemma 4B GGUF - Pan-African Languages

Quantized version of Google's TranslateGemma 4B for mobile deployment with llama.cpp.

Model Files

File Size Quantization Use Case
translategemma-4b-q2_k.gguf ~1.7GB Q2_K (2-bit) Budget Android phones (2-3GB RAM)

Supported Languages

  • West African: Twi, Ewe, Ga, Yoruba, Hausa, Igbo
  • East African: Swahili
  • Plus 15+ additional Pan-African languages

Usage with llama.cpp

./llama-cli -m translategemma-4b-q2_k.gguf -p "Translate to Twi: How are you feeling today?"

Part of Project Nku

This model is the "communicator" component of Nku - a medical triage system for West Africa.

  • MedGemma Q2_K: Medical reasoning (brains)
  • TranslateGemma Q2_K: Language translation (communicator)

Original Model

Based on google/translategemma-4b-it

Downloads last month
322
GGUF
Model size
3B params
Architecture
gemma
Hardware compatibility
Log In to add your hardware

1-bit

2-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support