Named Datasets

Practice and baselines need public hubs, not only private tables. This page is the hub list for public datasets and models, first for practice projects, then for translation, then for pre-trained models.

VARIOUS DATASETS

The quickest way to practice is a project that already comes with its data. 26 of them is Analytics Vidhya's list of machine learning projects for beginners, meant for learning data science by applying it and showing the skills on a CV.

Translation needs paired text rather than a single table. 24 is the Lionbridge list of parallel text datasets for machine translation training. For one concrete pair, English-Spanish, Eu-es is the Wolfram Data Repository's Europarl V7, a parallel corpus for machine translation from the proceedings of the European Parliament, and 2 is the same kind of resource on the European Data Portal.

When the baseline is a model rather than a dataset, the hub is 50K models deep: ModelDepot alone has over 50,000 freely accessible pre-trained models with search functionality to find them.