README.md

---
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
---

<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/jC7kdl8.jpeg" alt="TensorBlock" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
    <div style="display: flex; flex-direction: column; align-items: flex-start;">
        <p style="margin-top: 0.5em; margin-bottom: 0em;">
            Feedback and support: TensorBlock's  <a href="https://x.com/tensorblock_aoi">Twitter/X</a>, <a href="https://t.me/TensorBlock">Telegram Group</a> and <a href="https://x.com/tensorblock_aoi">Discord server</a>
        </p>
    </div>
</div>

## google/gemma-3-4b-it - GGUF

This repo contains GGUF format model files for [google/gemma-3-4b-it](https://huggingface.co/google/gemma-3-4b-it).

The files were quantized using machines provided by [TensorBlock](https://tensorblock.co/), and they are compatible with llama.cpp as of [commit b4882](https://github.com/ggml-org/llama.cpp/commit/be7c3034108473beda214fd1d7c98fd6a7a3bdf5).

## Our projects
<table border="1" cellspacing="0" cellpadding="10">
<tr>
  <th style="font-size: 25px;">Awesome MCP Servers</th>
  <th style="font-size: 25px;">TensorBlock Studio</th>
</tr>
  <tr>
    <th><img src="https://imgur.com/2Xov7B7.jpeg" alt="Project A" width="450"/></th>
    <th><img src="https://imgur.com/pJcmF5u.jpeg" alt="Project B" width="450"/></th>
  </tr>
  <tr>
    <th>A comprehensive collection of Model Context Protocol (MCP) servers.</th>
    <th>A lightweight, open, and extensible multi-LLM interaction studio.</th>
  </tr>
<tr>
  <th>
    <a href="https://github.com/TensorBlock/awesome-mcp-servers" target="_blank" style="
      display: inline-block;
      padding: 8px 16px;
      background-color: #FF7F50;
      color: white;
      text-decoration: none;
      border-radius: 6px;
      font-weight: bold;
      font-family: sans-serif;
    ">👀 See what we built 👀</a>
  </th>
  <th>
    <a href="https://github.com/TensorBlock/TensorBlock-Studio" target="_blank" style="
      display: inline-block;
      padding: 8px 16px;
      background-color: #FF7F50;
      color: white;
      text-decoration: none;
      border-radius: 6px;
      font-weight: bold;
      font-family: sans-serif;
    ">👀 See what we built 👀</a>
  </th>
</tr>
</table>
## Prompt template

```
<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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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](https://huggingface.co/tensorblock/gemma-3-4b-it-GGUF/blob/main/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

```shell
pip install -U "huggingface_hub[cli]"
```

Then, downoad the individual model file the a local directory

```shell
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:

```shell
huggingface-cli download tensorblock/gemma-3-4b-it-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf'
```