Question description
mission details
In recent years, with the development of artificial intelligence, it has achieved great success in many fields such as speech recognition, natural language processing, image and video analysis. With the government's call for environmental protection, garbage classification has become an urgent problem to be solved. This competition will focus on the classification of garbage pictures, using artificial intelligence technology to detect household garbage pictures and find out which categories are in the pictures. Rubbish. Participants are required to give an algorithm or model to detect the garbage category in the picture for a given picture. Given the picture data, the player trains the model based on this, and predicts the most correct category for each test data.
the data shows
The training and test pictures used in this competition are all from life scenes. There are a total of forty categories, and the correspondence between categories and labels is in the dict file in the training set. The category of garbage in the picture, the format is "Class 1/Class 2", and the second category is the specific garbage object category, which is the category marked in the training data, such as disposable fast food boxes, peels and pulp, old clothes, etc. There are four primary categories: recyclables, kitchen waste, hazardous waste and other waste.
The data files include training set (with labels) and test set (without labels). All pictures of the training set are stored in 0-39 folders under the train folder. The file name is the category label. The test set has 400 pieces to be The classified garbage pictures are in the test folder, and [login to view URL] saves the names of all test set files in the format: name+\n.
Submit answer
For exam submission, you need to submit the model code project version and result file . The result file is in TXT file format, named [login to view URL], and the fields in the file need to be written in the specified format.
The format of the submission result is as follows:
The number of rows in each category should correspond to the number of rows in the original data of the test set, and should not be out of order.
The output result should be checked for 400 rows of data, otherwise the result is invalid.
The output result file is named [login to view URL], one category label (number) per line
The sample is as follows:
···
35
3
2
37
10
3
26
4
34
twenty one
···