Instructions to use MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16
# Run inference directly in the terminal:
llama cli -hf MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16
# Run inference directly in the terminal:
llama cli -hf MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

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 MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16
# Run inference directly in the terminal:
./llama-cli -hf MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

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 MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

Use Docker

docker model run hf.co/MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

LM Studio
Jan

vLLM

How to use MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-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": "MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

Ollama
How to use MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF with Ollama:
```
ollama run hf.co/MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16
```
Unsloth Desktop

How to use MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF with Docker Model Runner:
```
docker model run hf.co/MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16
```

Lemonade

How to use MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

Run and chat with the model

lemonade run user.NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF-BF16

List all available models

lemonade list

Hermes Agent

How to use MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

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 MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16

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 "MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF:BF16" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

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

GGUFs made with a non-mainline build of llama.cpp: nvidia/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-BF16 (official), converted with convert_hf_to_gguf.py at Premshay/llama.cpp@731368b5d (branch nemotron-super-mtp, stacked on PR #25444), quantized with llama-quantize from the same build

The GGUFs should be loaded with the following fork of llama.cpp on the nemotron-super-mtp branch: https://github.com/Premshay/llama.cpp/tree/nemotron-super-mtp

Just an experiment to try the forked llama.cpp built on top of PR#25444 to support this model's built in MTP heads. Use the fork of llama.cpp above with --spec-type draft-mtp enabled in llama-server to provide a nice speedup in token generation.

The contributor of the fork explained their methodology on the model card below. I just did the same but with Nemotron Puzzle BF16 weights instead of using Nemotron Super: https://huggingface.co/Premshay/Nemotron-3-Super-120B-A12B-MTP-GGUF

Downloads last month: 521

GGUF

Model size

78B params

Architecture

nemotron_h_moe

Hardware compatibility

4-bit

5-bit

6-bit

16-bit

Model tree for MRockatansky/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-MTP-GGUF

Base model

nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16

Finetuned

nvidia/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-BF16

Quantized

(6)

this model