#Recent Changes
##H2O-Dev
###0.1.20.1019 - 1/19/15
####New Features
These features have been added since the last release:
#####UI
- Added various documentation links to the build page
#####API
- None
#####Algorithms
- Ported matrix multiply over and connected it to rapids
#####System
- None
####Enhancements
These changes are to improvements to existing features (which includes changed default values):
#####UI
- Allow user to specify (the log of) the number of rows per chunk for a new constant chunk; use this new function in CreateFrame
- Make CreateFrame non-blocking, now displays progress bar in Flow
- Add row and column count to H2OFrame show method
#####API
- Changed 2 to 3 for JSON requests
- Rename some more fields per consistency (
max_iterschanged tomax_iterations,_itersto_iterations,_ncatsto_categorical_column_count,_centersrawtocenters_raw,_avgwithinsstoavg_within_ss,_withinmsetowithin_mse, ) - Changed K-Means output parameters (
withinmsetowithin_mse,avgsstoavg_ss,avgbetweensstoavg_between_ss) - Remove default field values from DeepLearning parameters schema, since they come from the backing class
- Added ip_port field in node json output for Cloud query
#####Algorithms
- Minor fix in rapids matrix multiplicaton
- Updated sparse chunk to cut off binary search for prefix/suffix zeros
- Updated L_BFGS for GLM - warm-start solutions during lambda search, correctly pass current lambda value, added column-based gradient task
- Fix model parameters' default values in the metadata
- Set default value of k = number of clusters to 1 for K-Means
#####System
- Reject any training data with non-numeric values from KMeans model building
####Bug Fixes
These changes are to resolve incorrect software behavior:
#####UI
- None
#####API
- Fixed isSparse call for constant chunks
- Fixed sparse interface of constant chunks (no nonzero if const 1= 0)
#####Algorithms
- Fixed objective value in gradient solver (was missing penalty)
#####System
- Typeahead for folder contents apparently requires trailing "/"
- Fix build and instructions for R install.packages() style of installation; Note we only support source installs now
- Fixed R test runner h2o package install issue that caused it to fail to install on dev builds