Instructions to use NotMe404/gemma-4-31b-it-assistant-mtp-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NotMe404/gemma-4-31b-it-assistant-mtp-gguf with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="NotMe404/gemma-4-31b-it-assistant-mtp-gguf",
	filename="gemma4-31B-it-assistant-F16.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use NotMe404/gemma-4-31b-it-assistant-mtp-gguf with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M

Use Docker

docker model run hf.co/NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M

LM Studio
Jan
Ollama
How to use NotMe404/gemma-4-31b-it-assistant-mtp-gguf with Ollama:
```
ollama run hf.co/NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M
```

Unsloth Studio

How to use NotMe404/gemma-4-31b-it-assistant-mtp-gguf with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for NotMe404/gemma-4-31b-it-assistant-mtp-gguf to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for NotMe404/gemma-4-31b-it-assistant-mtp-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for NotMe404/gemma-4-31b-it-assistant-mtp-gguf to start chatting

Atomic Chat new
Docker Model Runner
How to use NotMe404/gemma-4-31b-it-assistant-mtp-gguf with Docker Model Runner:
```
docker model run hf.co/NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M
```

Lemonade

How to use NotMe404/gemma-4-31b-it-assistant-mtp-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull NotMe404/gemma-4-31b-it-assistant-mtp-gguf:Q4_K_M

Run and chat with the model

lemonade run user.gemma-4-31b-it-assistant-mtp-gguf-Q4_K_M

List all available models

lemonade list

Gemma 4 31B Assistant

The Gemma 4 31B it assistant converted to GGUF and is compatible with the latest llama.cpp release (b9549), which introduces support for Gemma 4 MTP (Multi-Token Prediction).

The GGUFs were produced using llama-quantize from the orginal Gemma 4 31b assistant https://huggingface.co/google/gemma-4-31B-it-assistant

Example usage

llama-server -m "gemma-4-31B-it-Q8_0.gguf" \
--spec-draft-model "gemma4-31B-it-assistant-Q8_0.gguf" \
--spec-type draft-mtp \
--spec-draft-n-max 4

Reccomended --spec-draft-n-max values:

--spec-draft-n-max 2
--spec-draft-n-max 3
--spec-draft-n-max 4

The results depend on your workload.

Multi-GPU example usage

It works fine with multiple gpus as well and I've seen a 2x increase in inference. The below example is for a 3 GPU setup.

llama-server -m "gemma-4-31B-it-Q8_0.gguf" \
--spec-draft-model "gemma4-31B-it-assistant-Q8_0.gguf" \
--spec-type draft-mtp \
--spec-draft-n-max 4 \
--main-gpu 0 \
--tensor-split 0.6,0.1,0.3

Downloads last month: 2,694

GGUF

Model size

0.5B params

Architecture

gemma4-assistant

Hardware compatibility

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for NotMe404/gemma-4-31b-it-assistant-mtp-gguf

Base model

google/gemma-4-31B-it-assistant

Quantized

(6)

this model