Commits : Listings

  Analyzed 3 days ago based on code collected 3 days ago.
Showing page 4 of 42
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Several minor updates to GUI so it looks acceptable under Linus Anon32 More... almost 2 years ago
Updated display of translation hypotheses to show each translation chunk and its alignment Anon32 More... almost 2 years ago
Modified calculation of alignment scores when there exist more than on sequence type. The default weight for Translation.Lexical.Number is as before but all other sequence types now default to zero. Anon32 More... almost 2 years ago
Fixed problem with gap-filling when two incomplete translations are adjacent. Anon32 More... almost 2 years ago
Fixed bug in FrequencyModel when handling gaps. Anon32 More... almost 2 years ago
Fixed threading issue in SumLogLinearModel Anon32 More... almost 2 years ago
Updated decoding to handle replacements Anon32 More... almost 2 years ago
More speed improvements when handling multiple sequence types. The largest bottleneck is now the FrequencyModel which can be quite slow when a translation includes a gap. Anon32 More... almost 2 years ago
Updated handling of multiple sequence types Anon32 More... almost 2 years ago
Better error checking for command arguments Anon32 More... almost 2 years ago
Simplified how multiple sequences and annotations could be specified on the command line. Added new LanguagePair class. The pipeline should now properly handle more than two languages, but this capability is not currently used. Anon32 More... almost 2 years ago
Divide n-gram f1 scores by n Anon32 More... almost 2 years ago
Changed default scaling factor in ReferenceMatcher as the log scores were causing underflow on very long sentences Anon32 More... almost 2 years ago
Several tweaks in decoding that should speed things up slightly. Added more features to SimilarityModel. Anon32 More... almost 2 years ago
Reference similarity shares the same code as target similarity. The target similarity computation should be faster as well. Anon32 More... almost 2 years ago
Added length ratio to similarity model Anon32 More... almost 2 years ago
Fixed bug where LogLinearModel.set() failed to update the score Anon32 More... almost 2 years ago
Updated target-side similarity scoring and added a few additional speed improvements. The translation scores changed slightly and this will require further investigation (they should not have changed). Anon32 More... almost 2 years ago
Removed cunei.options package and integrated the command line options with the cunei.config package. A few command line options have changed so look at the '-help' output if things do not behave as expected. All programs now allow the user to set arbitrary configuration options with the syntax '-Xoption=value'. Anon32 More... almost 2 years ago
Document context features were in the wrong order (they functioned correctly, but were effectively mislabeled). Anon32 More... almost 2 years ago
 
 
 

Creative Commons License Copyright © 2013 Black Duck Software, Inc. and its contributors, Some Rights Reserved. Unless otherwise marked, this work is licensed under a Creative Commons Attribution 3.0 Unported License . Ohloh ® and the Ohloh logo are trademarks of Black Duck Software, Inc. in the United States and/or other jurisdictions. All other trademarks are the property of their respective holders.