Instructions to use majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4 with MLX:

# Make sure mlx-vlm is installed
# pip install --upgrade mlx-vlm

from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
from mlx_vlm.utils import load_config

# Load the model
model, processor = load("majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4")
config = load_config("majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4")

# Prepare input
image = ["http://images.cocodataset.org/val2017/000000039769.jpg"]
prompt = "Describe this image."

# Apply chat template
formatted_prompt = apply_chat_template(
    processor, config, prompt, num_images=1
)

# Generate output
output = generate(model, processor, formatted_prompt, image)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4 with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4 with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "majentik/gemma-4-12B-it-TurboQuant-MLX-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 majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4

Run Hermes

hermes

OpenClaw new

How to use majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4 with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4"

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

gemma-4-12B-it — TurboQuant MLX NVFP4

google/gemma-4-12B-it @ 12ace6d648d72bd41519e140f1185f34d38c7e3d quantized pack, published as majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4.

Method

MLX quantization via mlx-vlm 0.6.3 (NVFP4, group_size 16); vision + audio towers retained in BF16 (not quantized).

Release line

Released under the TurboQuant line. RotorQuant and TurboQuant are this project's release labels for this pack, not distinct quantization algorithms — both brand repos for a given tier carry byte-identical weights, produced once and published under two names. No brand-specific speedup is claimed or measured for either label.

Modality

This pack is image-text-to-text capable: the vision and audio towers ship in BF16 alongside the quantized text tower, so image (and audio) inputs are supported end to end via mlx-vlm.

Vision fidelity note: for this instruct variant, vision has only been validated at 6-bit and 8-bit. Vision fidelity degrades below 6-bit; use 6-bit or 8-bit for image tasks.

License

Governed by the Gemma Terms of Use. See the upstream repo for the full license text.

Downloads last month: 82

Safetensors

Model size

3B params

Tensor type

U32

BF16

MLX

Hardware compatibility

4-bit

Model tree for majentik/gemma-4-12B-it-TurboQuant-MLX-NVFP4

Base model

google/gemma-4-12B

Finetuned

google/gemma-4-12B-it

Quantized

(288)

this model