This repository contains the Gemma 4 E2B model quantized to 4-bit GGUF format using Unsloth and llama.cpp. Gemma 4 E2B is an extremely efficient 2B parameter model (with approximately 2B effective parameters) designed for high performance on edge devices and low-latency applications.
Quantization Details
Quantization Format: GGUF (q4_k_m)
Quantization Method: llama.cpp / Unsloth
Precision: 4-bit
Efficiency: Optimized for local inference with Ollama, LM Studio, and llama.cpp.
Use with Ollama
You can run this model directly using Ollama:
ollama run hf.co/DuoNeural/Gemma-4-E2B-GGUF
Use with LM Studio
Open LM Studio.
Search for DuoNeural/Gemma-4-E2B-GGUF.
Download the Q4_K_M version and load it.
Architecture
Gemma 4 E2B is part of Google's latest lightweight model family, featuring state-of-the-art attention and architecture improvements that allow it to punch far above its weight class in coding and general reasoning.
Limitations
Performance may be limited for extremely long-form creative writing or highly complex multi-step logical puzzles compared to larger Gemma 4 variants.
Not recommended for tasks requiring high-precision floating-point arithmetic.
DuoNeural
DuoNeural is an open AI research lab — human + AI in collaboration.
