Gemma-4-E2B-NoAudio-GGUF

This repository provides GGUF weights for the text-only version of Gemma-4-E2B-it. The audio components have been removed for maximum efficiency on home hardware.

Key Features

Imatrix Optimized: All files are quantized using a custom Importance Matrix (imatrix) generated from 1,000+ chunks of Thai text to ensure superior reasoning and language quality.

Home Hardware Ready: Highly efficient and lightweight, perfect for running locally on consumer GPUs and CPUs.

Available Files

Q8_0 High fidelity for maximum accuracy.
Q4_K_M Best balance of speed and intelligence.

Usage

This model is compatible with most GGUF runners. For the best experience at home, we recommend:

LM Studio
Ollama
llama.cpp / llama-server

Downloads last month: 240

GGUF

Model size

5B params

Architecture

gemma4

Hardware compatibility

4-bit

Model tree for bombman/Gemma-4-E2B-NoAudio-GGUF

Base model

google/gemma-4-E2B-it

Quantized

(123)

this model