metadata
license: gemma
library_name: transformers
pipeline_tag: image-text-to-text
extra_gated_heading: Access Gemma on Hugging Face
extra_gated_prompt: >-
To access Gemma on Hugging Face, you’re required to review and agree to
Google’s usage license. To do this, please ensure you’re logged in to Hugging
Face and click below. Requests are processed immediately.
extra_gated_button_content: Acknowledge license
base_model: google/gemma-3-4b-it
tags:
- TensorBlock
- GGUF
Feedback and support: TensorBlock's Twitter/X, Telegram Group and Discord server
google/gemma-3-4b-it - GGUF
This repo contains GGUF format model files for google/gemma-3-4b-it.
The files were quantized using machines provided by TensorBlock, and they are compatible with llama.cpp as of commit b4882.
Our projects
| Awesome MCP Servers | TensorBlock Studio |
|---|---|
 |
 |
| A comprehensive collection of Model Context Protocol (MCP) servers. | A lightweight, open, and extensible multi-LLM interaction studio. |
| 👀 See what we built 👀 | 👀 See what we built 👀 |
<bos><start_of_turn>user
{system_prompt}
{prompt}<end_of_turn>
<start_of_turn>model
Model file specification
| Filename | Quant type | File Size | Description |
|---|---|---|---|
| gemma-3-4b-it-Q2_K.gguf | Q2_K | 1.729 GB | smallest, significant quality loss - not recommended for most purposes |
| gemma-3-4b-it-Q3_K_S.gguf | Q3_K_S | 1.937 GB | very small, high quality loss |
| gemma-3-4b-it-Q3_K_M.gguf | Q3_K_M | 2.098 GB | very small, high quality loss |
| gemma-3-4b-it-Q3_K_L.gguf | Q3_K_L | 2.236 GB | small, substantial quality loss |
| gemma-3-4b-it-Q4_0.gguf | Q4_0 | 2.363 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| gemma-3-4b-it-Q4_K_S.gguf | Q4_K_S | 2.378 GB | small, greater quality loss |
| gemma-3-4b-it-Q4_K_M.gguf | Q4_K_M | 2.490 GB | medium, balanced quality - recommended |
| gemma-3-4b-it-Q5_0.gguf | Q5_0 | 2.764 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| gemma-3-4b-it-Q5_K_S.gguf | Q5_K_S | 2.764 GB | large, low quality loss - recommended |
| gemma-3-4b-it-Q5_K_M.gguf | Q5_K_M | 2.830 GB | large, very low quality loss - recommended |
| gemma-3-4b-it-Q6_K.gguf | Q6_K | 3.191 GB | very large, extremely low quality loss |
| gemma-3-4b-it-Q8_0.gguf | Q8_0 | 4.130 GB | very large, extremely low quality loss - not recommended |
Downloading instruction
Command line
Firstly, install Huggingface Client
pip install -U "huggingface_hub[cli]"
Then, downoad the individual model file the a local directory
huggingface-cli download tensorblock/gemma-3-4b-it-GGUF --include "gemma-3-4b-it-Q2_K.gguf" --local-dir MY_LOCAL_DIR
If you wanna download multiple model files with a pattern (e.g., *Q4_K*gguf), you can try:
huggingface-cli download tensorblock/gemma-3-4b-it-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf'