Fwumious Wabbit is
- a very fast machine learning tool
- built with Rust
- inspired by and partially compatible with Vowpal Wabbit (much love! read more about compatibility here)
- currently supports logistic regression and (deep) field-aware factorization machines
Fwumious Wabbit is actively used in Outbrain for offline research, as well as for some production flows. It enables "high bandwidth research" when doing feature engineering, feature selection, hyperparameter tuning, and the like.
Data scientists can train hundreds of models over hundreds of millions of examples in a matter of hours on a single machine.
For our tested scenarios it is almost two orders of magnitude faster than the fastest Tensorflow implementation of Logistic Regression and FFMs that we could come up with. It is an order of magnitude faster than Vowpal Wabbit for some specific use-cases.
Check out our benchmark, here's a teaser:
Why is it faster? (see here for more details)
- Only implements Logistic Regression and (Deep) Field-aware Factorization Machines
- Uses hashing trick, lookup table for AdaGrad and a tight encoding format for the "input cache"
- Features' namespaces have to be declared up-front
- Prefetching of weights from memory (avoiding pipeline stalls)
- Written in Rust with heavy use of code specialization (via macros and traits)
- Special emphasis on efficiency of sparse operations and serving
This repo also contains the patching algorithm that enables very fast weight diff computation see weight_patcher for more details.