We ended up building a Time-Series Cross-Validation library from scratch
appliedexploration.comHi everyone,
We have been dealing for a while with the underperformance of Time-Series Machine Learning models (mostly due to regime changes), and haven't found the right library to complete Time Series Cross Validation before the heat-death of the universe.
We ended up writing a library from scratch, that comes with an order of magnitude speed-up, called Fold.
--- This is the launch of our core engine so we would love to get some feedback on Fold (https://github.com/dream-faster/fold)! We’ll be here and happy to answer any questions.
Does it work and have you tested it with PyTorch models?
Think: models that accept as input time-stamped multimodal data -- text, images, sound, etc.
We'll be working on the pytorch integration soon! `Fold`'s scope is time series, but there's nothing stopping you from sending in vector embeddings of any kind, timestamped.
Figuring out how to create those embeddings (that make sense over time) can mean quite a bit of research work, and requires flexibility, so it's probably better done outside of the time series library, with the tools of your choice.
> We'll be working on the pytorch integration soon! `Fold`'s scope is time series, but there's nothing stopping you from sending in vector embeddings of any kind, timestamped.
Awesome. I'll take a look. Thanks!
> Figuring out how to create those embeddings (that make sense over time) can mean quite a bit of research work, and requires flexibility, so it's probably better done outside of the time series library ...
Obviously :-)
Mark here (other co-founder), we're really curious if you are using Time Series Cross-Validation, with what tools, how frequently, and what kind of issues did you bump into!