grounded_sam

Grounding DINO

Input

• image

  (Image from https://github.com/IDEA-Research/Grounded-Segment-Anything/blob/main/assets/demo2.jpg)

• prompt

  The running dog.
  

Output

Requirements

This model requires additional module.

pip3 install ailia_tokenizer

Usage

Automatically downloads the onnx and prototxt files on the first run. It is necessary to be connected to the Internet while downloading.

For the sample image,

$ python3 grounded_sam.py

If you want to specify the input image, put the image path after the --input option.
You can use --savepath option to change the name of the output file to save.

$ python3 grounded_sam.py --input IMAGE_PATH --savepath SAVE_IMAGE_PATH

If you want to specify the caption for detection, put the caption after the --caption option. If you want to detect multiple objects in one sentence, we suggest separating each name with . . An example: cat. dog. chair.

$ python3 grounded_sam.py --caption "The running dog."

Reference

• Grounded-SAM

Framework

Pytorch

Model Format

ONNX opset=17

Netron

sam_vit_h_4b8939.onnx.prototxt