Instructions to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF", dtype="auto")

llama-cpp-python

How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF",
	filename="gemma-4-31B-it-NVFP4-turbo-NVFP4.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4
# Run inference directly in the terminal:
llama-cli -hf CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4
# Run inference directly in the terminal:
llama-cli -hf CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

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 CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4
# Run inference directly in the terminal:
./llama-cli -hf CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

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 CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4
# Run inference directly in the terminal:
./build/bin/llama-cli -hf CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

Use Docker

docker model run hf.co/CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

LM Studio
Jan

vLLM

How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

SGLang

How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with Ollama:
```
ollama run hf.co/CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4
```

Unsloth Studio

How to use CISCai/gemma-4-31B-it-NVFP4-turbo-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 CISCai/gemma-4-31B-it-NVFP4-turbo-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 CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF to start chatting

How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

Run Hermes

hermes

Docker Model Runner
How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with Docker Model Runner:
```
docker model run hf.co/CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4
```

Lemonade

How to use CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF:NVFP4

Run and chat with the model

lemonade run user.gemma-4-31B-it-NVFP4-turbo-GGUF-NVFP4

List all available models

lemonade list

⚡ Gemma 4 31B IT NVFP4 Turbo GGUF

Requires ggml-org/llama.cpp#21971

A repackaged nvidia/Gemma-4-31B-IT-NVFP4 that is 68% smaller in GPU memory and ~2.5× faster than the base model, while retaining nearly identical quality (1-3% loss). Fits on a single RTX 5090 (🎉).

Approach

Three changes were made:

Quantized all self-attention weights from BF16 → FP4 (RTN, group_size=16, matching modelopt NVFP4 format)
Updated architecture to Gemma4ForCausalLM and quantization config accordingly
Stripped the vision and audio encoder

Everything else is untouched — MLP layers keep NVIDIA's calibrated FP4, embed_tokens stays BF16, all norms preserved, so we retain all the nvidia/Gemma-4-31B-IT-NVFP4 optimizations.

Why RTN didn't hurt quality

RTN (Round-To-Nearest) is the simplest quantization method — no calibration data, fully reproducible. It worked here because:

FP4 with group_size=16 and per-group scaling preserves relative weight distributions well
Self-attention weights tend to be normally distributed near zero, where the FP4 grid has finest resolution (0, 0.5, 1.0, 1.5)
MLP layers (more sensitive to quantization) keep NVIDIA's calibrated FP4
embed_tokens stays BF16, preventing noise from propagating through all layers

License

Apache 2.0 — same as the base model.

Credits

Google DeepMind for Gemma 4
NVIDIA for the modelopt NVFP4 checkpoint

Downloads last month: 13,453

GGUF

Model size

31B params

Architecture

gemma4

Hardware compatibility

4-bit

Model tree for CISCai/gemma-4-31B-it-NVFP4-turbo-GGUF

Base model

google/gemma-4-31B

Finetuned

google/gemma-4-31B-it

Quantized

LilaRest/gemma-4-31B-it-NVFP4-turbo

Quantized

(1)

this model

Evaluation results

Accuracy on GPQA Diamond
self-reported

72.730
Accuracy on MMLU Pro
self-reported

83.930