smartdolphin / shopee

Shopee - Price Match Guarantee (kaggle challenge 2021)

Shopee - Price Match Guarantee

Shopee - Price Match Guarantee challenge is to find similar images for each group.
I participated in this challenge from 2021-05-03 to 2021-05-11.

Summary of approach

Private score: ? / Public score: 0.734 (entered top 9%)

Architecture

• nfnet-l0 + efficientnet-b5 + tfidf + transformer (arcface)
• threshold finding
• feature merge strategy

Data

the training set metadata. Each row contains the data for a single posting. Multiple postings might have the exact same image ID, but with different titles or vice versa.

• Num. of train : 34,250
• Num. of test : 70,000 (only the first few rows/images of the test set are published)

unzip shopee-product-matching.zip to local.

└── train.csv
└── test.csv
└── sample_submission.csv
└── train_images
    └── 55a8b996ee39086107e141014d5e651f.jpg
    └── ab8c2ccda452af172481842c1dd3b3d7.jpg
    └── ...
└── test_images
    └── 0006c8e5462ae52167402bac1c2e916e.jpg
    └── 0007585c4d0f932859339129f709bfdc.jpg
    └── 0008377d3662e83ef44e1881af38b879.jpg    

Feature

• posting_id - the ID code for the posting.
• image - the image id/md5sum.
• image_phash - a perceptual hash of the image.
• title - the product description for the posting.
• label_group - ID code for all postings that map to the same product. Not provided for the test set.

Sumission

• posting_id - the ID code for the posting.
• matches - Space delimited list of all posting IDs that match this posting. Posts always self-match. Group sizes were capped at 50, so there's no need to predict more than 50 matches.

Metric

• Metric will be evaluated based on their mean F1 score.