pietrolesci/amazoncat-13k

This dataset has been created as an artefact of the paper AnchorAL: Computationally Efficient Active Learning for Large and Imbalanced Datasets (Lesci and Vlachos, 2024). More info about this dataset in the appendix of the paper.

The AmazonCat-13k dataset was released by McAuley and Leskovec (2013) and is composed of product descriptions and reviews classified into 13k multi-label categories. The dataset is split into 1.2M train and 300k evaluation instances. It is commonly used as an extreme classification benchmark (You et al., 2019) where the goal is to classify an item into its categories. The data are exactly the same as the original available at this Google Drive link with the following additions:

1. 1.A unique identifier, uid column.

1. 1.Indices, that is, 3 columns with the embeddings of 3 different sentence-transformers
2. 2.all-mpnet-base-v2
3. 3.multi-qa-mpnet-base-dot-v1
4. 4.all-MiniLM-L12-v2

1. 1.Renaming of the label column to labels for easier compatibility with the transformers library.