Age | Commit message (Expand) | Author |
2012-03-20 | enable multiple cores per fold | Eva |
2012-03-19 | update training script to enable multi-fold training | Eva Hasler |
2012-03-19 | change param mixWithoutAveraging to --sparse-no-average, enable training on m... | Eva |
2012-03-18 | add parameters --scale-margin-precision, --scale-update-precision | Eva |
2012-03-15 | fix history Bleu and add parameter --use-source-length-history | Eva Hasler |
2012-03-14 | add parameter --mix-without-averaging | Eva Hasler |
2012-03-13 | replace parameter --mix-by-averaging with --sparse-average, fix bug in shardi... | Eva |
2012-03-13 | add parameter --dump-mixed-weights | Eva |
2012-03-13 | fix: mixing without averaging | Eva |
2012-03-12 | change history Bleu, add scaling option for sentence Bleu | Eva |
2012-03-12 | add parameter --mix-by-averaging | Eva |
2012-03-12 | update partial updates before next update | Eva |
2012-03-11 | add parameter --batch-equals-shard, --separate-updates | Eva Hasler |
2012-03-08 | tempor. change margin normalisation to normalise loss and modelScoreDiff | Eva Hasler |
2012-03-07 | Merge branch 'miramerge' of thor.inf.ed.ac.uk:/fs/saxnot3/ehasler/mosesdecode... | Eva Hasler |
2012-03-07 | add option --scale-by-inverse-length | Eva Hasler |
2012-03-07 | enable mpi for hope/fear/model decoding | Eva |
2012-03-06 | add parameter --normalise-margin | Eva Hasler |
2012-03-05 | fix decode-hope/decode-fear | Eva Hasler |
2012-03-05 | add schemes 6, 7 | Eva Hasler |
2012-03-05 | add parameter --hope-model | Eva Hasler |
2012-03-05 | add scheme 6 | Eva Hasler |
2012-03-05 | specify 5 different update schemes | Eva |
2012-03-04 | print features parameters --decode-hope, --decode-fear | Eva |
2012-03-02 | Merge branch 'miramerge' of thor.inf.ed.ac.uk:/fs/saxnot3/ehasler/mosesdecode... | Eva Hasler |
2012-03-02 | add option --decode-model | Eva Hasler |
2012-03-01 | fix length check | Eva Hasler |
2012-03-01 | enable nbest hope/fear decoding | Eva Hasler |
2012-03-01 | change hope decoding output | Eva Hasler |
2012-03-01 | Merge branch 'miramerge' of thor.inf.ed.ac.uk:/fs/saxnot3/ehasler/mosesdecode... | Eva Hasler |
2012-03-01 | dont print out weight vectors | Eva Hasler |
2012-03-01 | add params --decode-hope, --decode-fear | Eva Hasler |
2012-02-29 | dont print out weight vectors | Eva Hasler |
2012-02-21 | remove warning.. | Eva Hasler |
2012-02-21 | add modes 'rank-only' and 'hope-fear-rank' | Eva Hasler |
2012-02-20 | compute model score difference from model scores instead of from feature valu... | Eva Hasler |
2012-02-20 | introduce boost parameter | Eva Hasler |
2012-02-19 | enable mira learning rate for analytical update | Eva Hasler |
2012-02-15 | separate decoder settings for training | Eva Hasler |
2012-02-14 | remove output of feature value difference | Eva Hasler |
2012-02-14 | remove output of feature value difference | Eva Hasler |
2012-02-14 | remove -mbr parameter from Decoder.cpp | Eva Hasler |
2012-02-14 | fix problem with glm feature names | Eva Hasler |
2012-02-04 | remove correct-scaling vars | Eva HASLER |
2012-02-04 | Merge branch 'miramerge' of thor.inf.ed.ac.uk:/fs/saxnot3/ehasler/mosesdecode... | Eva HASLER |
2012-02-04 | remove --correct-scaling, introduce --min-bleu-ratio | Eva HASLER |
2012-02-02 | ignore details | Eva Hasler |
2012-02-01 | fix --core-weights, remove stabilisation code | Eva Hasler |
2012-02-01 | Merge remote branch 'github/miramerge' into bjam | Barry Haddow |
2012-01-31 | fix scaling correction.. (hopefully) | Eva Hasler |