Image Captioning with ClipCap

author: David Wang

Description

This operator generates the caption with ClipCap which describes the content of the given image. ClipCap uses CLIP encoding as a prefix to the caption, by employing a simple mapping network, and then fine-tunes a language model to generate the image captions. This is an adaptation from rmokady/CLIP_prefix_caption.

Code Example

Load an image from path './hulk.jpg' to generate the caption.

Write the pipeline in simplified style:

import towhee

towhee.glob('./hulk.jpg') \
      .image_decode() \
      .image_captioning.clipcap(model_name='clipcap_coco') \
      .show()

Write a same pipeline with explicit inputs/outputs name specifications:

import towhee

towhee.glob['path']('./hulk.jpg') \
      .image_decode['path', 'img']() \
      .image_captioning.clipcap['img', 'text'](model_name='clipcap_coco') \
      .select['img', 'text']() \
      .show()

Factory Constructor

Create the operator via the following factory method

clipcap(model_name)

Parameters:

​ model_name: str

​ The model name of ClipCap. Supported model names:

• clipcap_coco
• clipcap_conceptual

Interface

An image-text embedding operator takes a towhee image as input and generate the correspoing caption.

Parameters:

​ data: towhee.types.Image (a sub-class of numpy.ndarray)

​ The image to generate embedding.

Returns: str

​ The caption generated by model.