Method for detecting defect and method for training model

A method for determining a target model comprising:

• acquiring, in response to a configuration operation of a user on a sample data set, the sample data set, the sample data set comprising a plurality of sample images that have been labeled with a category, and the sample data set comprising a training data set and a verification data set;
• configuring a training parameter according to feature information of the sample data set, and generating and displaying a training parameter display interface, the training parameter displayed in the training parameter display interface comprising a test strategy, the test strategy comprising at least two numbers of training rounds, and the feature information comprising a number of samples in the sample data set;
• training, according to the training parameter, a same deep learning model by inputting the training data set into the same deep learning model to obtain at least two trained models, the at least two trained models corresponding to the at least two numbers of training rounds respectively;
• testing the at least two trained models by inputting the verification data set into the at least two trained models respectively to generate a verification test result;
• generating and displaying, based on the verification test result, a verification test index display interface, the verification test index display interface being configured to display a confusion matrix, an accuracy rate, a recall rate and an F1 score;
• and
• determining, in response to a selection operation by the user for the at least two trained models, a model selected by the user as the target model.

A computing device, comprising:

• a processor;
• and
• a memory, storing computer-executable instructions that, when being executed by the processor, cause the computing device to perform:
• acquiring a sample data set, the sample data set comprising a plurality of sample images that have been labeled with a category, and the sample data set comprising a training data set and a verification data set;
• training a same deep learning model by inputting the training data set into the same deep learning model to obtain at least two trained models, wherein the at least two trained models correspond to different numbers of training rounds of the same deep learning model;
• testing the at least two trained models by inputting the verification data set into the at least two trained models respectively to generate a verification test result;
• generating, based on the verification test result, a verification test index, the verification test index comprising a confusion matrix, an accuracy rate, a recall rate and an F1 score;
• generating and displaying a verification test index display interface configured to display the confusion matrix, the accuracy rate, the recall rate and the F1 score of the verification test index;
• and
• determining, according to the verification test index, the target model from the at least two trained models.