VLN-1-Mini-3B-Lite
VLN-1-Mini-3B-Lite is a fast, lightweight 3-billion-parameter language model developed and fine-tuned by Nehonex for on-device and local inference.
Model Details
- Developer: Nehonex
- Model Family: VLN-1
- Parameters: 3 Billion (4-bit Q4_K_M quantized, ~1.93 GB)
- Use Cases: Mobile offline assistant, coding help, general Q&A
Quick Start (Ollama)
ollama run hf.co/Nehonex/VLN-1-Mini-3B-Lite:Q4_K_M
License & Attribution
Released under the Apache 2.0 License. Fine-tuned from the Qwen2.5-3B-Instruct architecture.