by melamed
- July 25, 2014, 23:08:32 CET
[
]
18650 views,
5360 downloads, 0 subscriptions
|
About:
Boosting algorithms for classification and regression, with many variations. Features include:
Scalable and robust;
Easily customizable loss functions;
One-shot training for an entire regularization path;
Continuous checkpointing;
much more
Changes:
-
added ElasticNets as a regularization option
-
fixed some segfaults, memory leaks, and out-of-range errors, which were creeping in in some corner cases
-
added a couple of I/O optimizations
|