Welcome to
mirror list
, hosted at
ThFree Co
, Russian Federation.
github.com/moses-smt/nplm.git - Unnamed repository; edit this file 'description' to name the repository.
index
:
github.com/moses-smt/nplm.git
master
Unnamed repository; edit this file 'description' to name the repository.
www-data
summary
refs
log
tree
commit
diff
log msg
author
committer
range
Age
Commit message (
Expand
)
Author
2015-09-03
unbork commit 9dea3f
HEAD
master
Rico Sennrich
2015-08-27
(optionally) only save best model (based on validation perplexity)
Rico Sennrich
2015-08-27
Makefile option (-DNPLM_DOUBLE_PRECISION) to switch between double and float.
Rico Sennrich
2015-08-27
speed up createMmap
Rico Sennrich
2015-08-27
input dropout
Rico Sennrich
2015-07-20
test commit
Hieu Hoang
2015-07-17
Merge pull request #5 from graehl/master
Rico Sennrich
2015-06-26
find_string
graehl
2015-06-25
lookup_word nplm 0.1 backward compat
graehl
2015-06-25
tab fix
graehl
2015-06-25
compile
graehl
2015-06-25
vocabulary no-copy find_string pair-char-ptrs
graehl
2015-06-25
more tab->space
graehl
2015-06-25
fix mixed tab2/tab4/spaces indents
graehl
2015-06-25
compile warnings
graehl
2015-06-25
c++11 compile
graehl
2015-04-08
refactor validation perplexity test; check perplexity every 500000 minibatches.
Rico Sennrich
2015-04-08
options for premultiplication and unnormalized networks in testNeuralNetwork
Rico Sennrich
2015-04-07
training instances in last minibatch should not get higher weight
Rico Sennrich
2015-03-25
Merge pull request #4 from KentonMurray/master
Rico Sennrich
2015-03-25
Many of the header files contained an absolute path to Eigen. I have fixed th...
Kenton William Murray
2015-02-24
switch on more verbose output for testNeuralNetwork with option '--debug 1'
Rico Sennrich
2015-02-19
bugfix: get_output_vocabulary actually returns input vocabulary
Rico Sennrich
2015-02-12
bugfix: don't overwrite input/output vocabulary sizes if vocab size is 0.
Rico Sennrich
2015-02-11
be less verbose
Rico Sennrich
2015-02-11
createMmap: convert ngramized and numberized text files into memory mapped fi...
Rico Sennrich
2014-11-17
python: support hidden biases
Rico Sennrich
2014-11-17
re-apply 64120 (support for networks with single hidden layer)
Rico Sennrich
2014-11-17
re-apply cfc957 (expose vocabulary map to Moses)
Rico Sennrich
2014-11-17
re-apply 233de0 (Eigen library assumed to be in 3rdparty directory)
Rico Sennrich
2014-11-17
re-apply 31412f (osx compile)
Rico Sennrich
2014-11-17
cleanup Makefile
Rico Sennrich
2014-11-17
expose premultiplication
Rico Sennrich
2014-11-17
re-apply ba48d (python: distinct input/output vocabs; optionally skip one hid...
Rico Sennrich
2014-11-17
upstream 0.3 (reverting all changes from this fork)
Rico Sennrich
2014-11-04
python: distinct input/output vocabs; optionally skip one hidden layer
Rico Sennrich
2014-10-16
Expose the vocabulary map to use with Moses
XapaJIaMnu
2014-10-16
support for networks with single hidden layer
Rico Sennrich
2014-10-16
new parameter init_model to initialize the network parameters from an existin...
Rico Sennrich
2014-10-16
Revert shared cache. Doesn't save enough memory to be worth locking the syste...
Rico Sennrich
2014-09-02
no longer need to put Eigen include in system directory. Need to put it in np...
Hieu Hoang
2014-08-28
Merge pull request #1 from hieuhoang/master
Rico Sennrich
2014-08-28
osx compile
Hieu Hoang
2014-08-28
osx compile
Hieu Hoang
2014-07-09
shared cache (thread-safe with WITH_THREADS)
Rico Sennrich
2014-07-04
allow premultiplication during initialization
Rico Sennrich
2013-11-08
Const kludge
Kenneth Heafield
2013-11-05
Generate libneuralLM.so instead of neuralLM.so. Same for .a
Kenneth Heafield
2013-11-05
Pull out my own method call for queries
Kenneth Heafield
2013-10-30
Add .gitignore
Kenneth Heafield
[next]