Instructions to use dill-dev/NanoDream-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use dill-dev/NanoDream-7B 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 dill-dev/NanoDream-7B:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf dill-dev/NanoDream-7B:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf dill-dev/NanoDream-7B:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf dill-dev/NanoDream-7B: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 dill-dev/NanoDream-7B:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf dill-dev/NanoDream-7B: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 dill-dev/NanoDream-7B:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf dill-dev/NanoDream-7B:Q4_K_M

Use Docker

docker model run hf.co/dill-dev/NanoDream-7B:Q4_K_M

LM Studio
Jan

vLLM

How to use dill-dev/NanoDream-7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dill-dev/NanoDream-7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dill-dev/NanoDream-7B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/dill-dev/NanoDream-7B:Q4_K_M

Ollama
How to use dill-dev/NanoDream-7B with Ollama:
```
ollama run hf.co/dill-dev/NanoDream-7B:Q4_K_M
```

Unsloth Studio

How to use dill-dev/NanoDream-7B 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 dill-dev/NanoDream-7B 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 dill-dev/NanoDream-7B to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for dill-dev/NanoDream-7B to start chatting

Docker Model Runner
How to use dill-dev/NanoDream-7B with Docker Model Runner:
```
docker model run hf.co/dill-dev/NanoDream-7B:Q4_K_M
```

Lemonade

How to use dill-dev/NanoDream-7B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull dill-dev/NanoDream-7B:Q4_K_M

Run and chat with the model

lemonade run user.NanoDream-7B-Q4_K_M

List all available models

lemonade list

Atomic Chat

🎨 NanoDream-7B (GGUF)

NanoDream-7B is a high-performance, next-generation multimodal model optimized for efficiency, speed, and advanced image reasoning. This model brings professional-grade Vision-Language capabilities to consumer-grade hardware, laptops, and mobile devices using the GGUF format.

🚀 Key Highlights

Optimized Architecture: Fine-tuned for high-speed multi-modal reasoning.
Quantization: Q4_K_M (The industry standard for balancing quality and performance).
Low Resource Usage: Runs comfortably on devices with 8GB RAM or less.
Unified Interface: Perfect for real-time image description, object detection, and visual QA.

🛠️ Quantization Details

This model was quantized using llama.cpp to provide a seamless experience on local hardware.

Method: Q4_K_M (4-bit quantization with medium-sized K-quants)
Format: GGUF (Compatible with llama.cpp, LM Studio, and more)
Model Size: Approx. 4.08 GB

💻 How to Use

1. Using llama.cpp (Command Line)

To interact with NanoDream-7B via terminal, use the following command:

./llama-cli \
  -m NanoDream-7B-Q4_K_M.gguf \
  --mmproj NanoDream-7B-mmproj-f16.gguf \
  --image input_sample.jpg \
  -p "Describe this image accurately."

2. Prompt Template

For best results, use the standard interaction format:

USER: <image>\n<prompt>\nASSISTANT:

📊 Hardware Requirements

Resource	Minimum	Recommended
System RAM	6 GB	8 GB+
VRAM (GPU)	4 GB	6 GB+
Disk Space	4.5 GB	5 GB

🛡️ Disclaimer

NanoDream-7B is a powerful tool for visual understanding. However, users should verify critical information generated by the model. It is not intended for use in high-risk medical, legal, or safety-critical applications.

Maintained and Published by: dill-dev

Downloads last month: 13

GGUF

Model size

7B params

Architecture

llama

Hardware compatibility

4-bit