Large model prompt learning method based on category attribute knowledge enhancement

The invention discloses a large model prompt learning method based on category attribute knowledge enhancement, and the method comprises the steps: obtaining an image recognition training data set, and generating a visual attribute set of each category through manual annotation or the use of ChatGPT; generating an attribute perceptible prompt background through an attribute integration module; respectively putting the training images and the corresponding prompt sentences carrying the visual attribute information into an image encoder and a text encoder to obtain image features and text classification weights; performing comparative learning on the image features and the text classification weights, and updating parameters of an attribute integration module through comparative learning to obtain a trained attribute integration module; generating a most significant visual attribute set of each test category according to the category space of the test task; and loading a to-be-tested image and the visual attrib