3 ms·
Show HN: Training synthetic models on highly complex datasets
- repeat_or 5y agoWe published a notebook and a GitHub repo that helps you train synthetic models on highly dimensional datasets (e.g. 1000's of columns, and millions of records). It works by using Gretel's open source header clustering to group correlated data and parallelize training across multiple GPUs.
- ag408 5y agoThanks for posting this :)