blip2

BLIP-2

Input

(Image from https://github.com/salesforce/LAVIS/blob/main/docs/_static/merlion.png)

Output

• Estimating Caption

### Caption ###
singapore merlion fountain

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 blip2.py

If you want to specify the input image, put the image path after the --input option.

$ python3 blip2.py --input IMAGE_PATH

Reference

• LAVIS - BLIP-2
• Hugging Face - BLIP-2

Framework

Pytorch

Model Format

ONNX opset=11

Netron

blip2-opt-2.7b.onnx.prototxt
vision_model.onnx.prototxt