The solution for competition: Steel-Defect-Detection

This is a contest from kaggle, part of the kernel comes from the notebook interface of the contest inferface, and its corresponding link will be placed at the begining of each notebook's article.

Here is the competition link: https://www.kaggle.com/c/severstal-steel-defect-detection

The document structure is as follows:

• input
  • train_images
  • test_images
  • sample_submission.csv
  • train.csv
• visualize_data.ipynb