/moondream

tiny vision language model

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

🌔 moondream

a tiny vision language model that kicks ass and runs anywhere

Website | Hugging Face | Demo

Benchmarks

moondream2 is a 1.86B parameter model initialized with weights from SigLIP and Phi 1.5.

Model VQAv2 GQA TextVQA TallyQA (simple) TallyQA (full)
moondream1 74.7 57.9 35.6 - -
moondream2 (latest) 77.7 61.7 49.7 80.1 74.2

Examples

Image Example
What is the girl doing?
The girl is sitting at a table and eating a large hamburger.

What color is the girl's hair?
The girl's hair is white.
What is this?
The image features a computer server rack, which is a large metal structure designed to hold and organize multiple computer components, such as motherboards, cooling systems, and other peripherals. The rack is filled with various computer parts, including multiple computer chips, wires, and other electronic components. The rack is placed on a carpeted floor, and there is a couch in the background, suggesting that the setup is likely in a living or working space.

What is behind the stand?
There is a brick wall behind the stand.

Usage

Using transformers (recommended)

pip install transformers einops
from transformers import AutoModelForCausalLM, AutoTokenizer
from PIL import Image

model_id = "vikhyatk/moondream2"
revision = "2024-04-02"
model = AutoModelForCausalLM.from_pretrained(
    model_id, trust_remote_code=True, revision=revision
)
tokenizer = AutoTokenizer.from_pretrained(model_id, revision=revision)

image = Image.open('<IMAGE_PATH>')
enc_image = model.encode_image(image)
print(model.answer_question(enc_image, "Describe this image.", tokenizer))

The model is updated regularly, so we recommend pinning the model version to a specific release as shown above.

To enable Flash Attention on the text model, pass in attn_implementation="flash_attention_2" when instantiating the model.

model = AutoModelForCausalLM.from_pretrained(
    model_id, trust_remote_code=True, revision=revision,
    torch_dtype=torch.float16, attn_implementation="flash_attention_2"
).to("cuda")

Batch inference is also supported.

answers = moondream.batch_answer(
    images=[Image.open('<IMAGE_PATH_1>'), Image.open('<IMAGE_PATH_2>')],
    prompts=["Describe this image.", "Are there people in this image?"],
    tokenizer=tokenizer,
)

Using this repository

Clone this repository and install dependencies.

pip install -r requirements.txt

sample.py provides a CLI interface for running the model. When the --prompt argument is not provided, the script will allow you to ask questions interactively.

python sample.py --image [IMAGE_PATH] --prompt [PROMPT]

Use gradio_demo.py script to start a Gradio interface for the model.

python gradio_demo.py

webcam_gradio_demo.py provides a Gradio interface for the model that uses your webcam as input and performs inference in real-time.

python webcam_gradio_demo.py

Limitations

  • The model may generate inaccurate statements, and struggle to understand intricate or nuanced instructions.
  • The model may not be free from societal biases. Users should be aware of this and exercise caution and critical thinking when using the model.
  • The model may generate offensive, inappropriate, or hurtful content if it is prompted to do so.