diff options
| -rw-r--r-- | training/crf/mpi_online_optimize.cc | 12 | ||||
| -rwxr-xr-x | word-aligner/aligner.pl | 19 | ||||
| -rw-r--r-- | word-aligner/makefiles/makefile.grammars | 2 | ||||
| -rw-r--r-- | word-aligner/makefiles/makefile.model.f-e | 14 |
4 files changed, 44 insertions, 3 deletions
diff --git a/training/crf/mpi_online_optimize.cc b/training/crf/mpi_online_optimize.cc index d6968848..9e1ae34c 100644 --- a/training/crf/mpi_online_optimize.cc +++ b/training/crf/mpi_online_optimize.cc @@ -5,6 +5,7 @@ #include <cassert> #include <cmath> #include <tr1/memory> +#include <ctime> #include <boost/program_options.hpp> #include <boost/program_options/variables_map.hpp> @@ -41,6 +42,7 @@ bool InitCommandLine(int argc, char** argv, po::variables_map* conf) { ("training_agenda,a",po::value<string>(), "Text file listing a series of configuration files and the number of iterations to train using each configuration successively") ("minibatch_size_per_proc,s", po::value<unsigned>()->default_value(5), "Number of training instances evaluated per processor in each minibatch") ("optimization_method,m", po::value<string>()->default_value("sgd"), "Optimization method (sgd)") + ("max_walltime", po::value<unsigned>(), "Maximum walltime to run (in minutes)") ("random_seed,S", po::value<uint32_t>(), "Random seed (if not specified, /dev/random will be used)") ("eta_0,e", po::value<double>()->default_value(0.2), "Initial learning rate for SGD (eta_0)") ("L1,1","Use L1 regularization") @@ -304,6 +306,9 @@ int main(int argc, char** argv) { int write_weights_every_ith = 100; // TODO configure int titer = -1; + unsigned timeout = 0; + if (conf.count("max_walltime")) timeout = 60 * conf["max_walltime"].as<unsigned>(); + const time_t start_time = time(NULL); for (int ai = 0; ai < agenda.size(); ++ai) { const string& cur_config = agenda[ai].first; const unsigned max_iteration = agenda[ai].second; @@ -336,9 +341,14 @@ int main(int argc, char** argv) { ostringstream o; o << "weights.epoch_" << (ai+1) << '.' << iter << ".gz"; fname = o.str(); } + const time_t cur_time = time(NULL); + if (timeout) { + if ((cur_time - start_time) > timeout) converged = true; + } if (converged && ((ai+1)==agenda.size())) { fname = "weights.final.gz"; } ostringstream vv; - vv << "total iter=" << titer << " (of current config iter=" << iter << ") minibatch=" << size_per_proc << " sentences/proc x " << size << " procs. num_feats=" << x.size() << '/' << FD::NumFeats() << " passes_thru_data=" << (titer * size_per_proc / static_cast<double>(corpus.size())) << " eta=" << lr->eta(titer); + double minutes = (cur_time - start_time) / 60.0; + vv << "total walltime=" << minutes << "min iter=" << titer << " (of current config iter=" << iter << ") minibatch=" << size_per_proc << " sentences/proc x " << size << " procs. num_feats=" << x.size() << '/' << FD::NumFeats() << " passes_thru_data=" << (titer * size_per_proc / static_cast<double>(corpus.size())) << " eta=" << lr->eta(titer); const string svv = vv.str(); cerr << svv << endl; Weights::WriteToFile(fname, lambdas, true, &svv); diff --git a/word-aligner/aligner.pl b/word-aligner/aligner.pl index c5078645..cbccb94a 100755 --- a/word-aligner/aligner.pl +++ b/word-aligner/aligner.pl @@ -51,6 +51,8 @@ while(<IN>) { chomp; my ($f, $e) = split / \|\|\| /; die "Bad format, excepted ||| separated line" unless defined $f && defined $e; + $f =~ s/\[/(/g; + $e =~ s/\]/)/g; print F "$f\n"; print E "$e\n"; } @@ -80,6 +82,11 @@ NCLASSES = $num_classes TARGETS = @targets PTRAIN = \$(TRAINING_DIR)/cluster-ptrain.pl --restart_if_necessary PTRAIN_PARAMS = --gaussian_prior --sigma_squared 1.0 --max_iteration 15 +#MPIJOBS = 4 +#MPIRUN = mpirun -np $(MPIJOBS) +MPIRUN= + +WALLTIME=90 export @@ -99,7 +106,15 @@ clean: EOT close TOPLEVEL; -print STDERR "Created alignment task. chdir to talign/ then type make.\n\n"; +print STDERR <<EOT; +Created alignment task. To start, run: +cd talign/ +make + +To specify the walltime *in minutes* used by the optimizer, use +make WALLTIME=120 + +EOT exit 0; sub make_stage { @@ -142,6 +157,8 @@ EOT open AGENDA, ">$stage_dir/agenda.txt" or die "Can't write $stage_dir/agenda.txt: $!"; print AGENDA "cdec.ini $TRAINING_ITERATIONS\n"; close AGENDA; + `cp $SCRIPT_DIR/makefiles/makefile.model.$direction $stage_dir/Makefile`; + die unless $? == 0; } sub usage { diff --git a/word-aligner/makefiles/makefile.grammars b/word-aligner/makefiles/makefile.grammars index ce3e1638..8d3ea8cb 100644 --- a/word-aligner/makefiles/makefile.grammars +++ b/word-aligner/makefiles/makefile.grammars @@ -4,7 +4,7 @@ clean: $(RM) orthonorm-dict.* voc2class* corpus.class.* corpus.e-f corpus.f-e corpus.f-e.lex-grammar* *.model1 *voc corpus.e-f.lex-grammar* *stem* freq* wordpairs* SUPPORT_DIR = $(SCRIPT_DIR)/support -GZIP = /usr/bin/gzip +GZIP = gzip ZCAT = zcat EXTRACT_GRAMMAR = $(SUPPORT_DIR)/extract_grammar.pl EXTRACT_VOCAB = $(SUPPORT_DIR)/extract_vocab.pl diff --git a/word-aligner/makefiles/makefile.model.f-e b/word-aligner/makefiles/makefile.model.f-e new file mode 100644 index 00000000..404f5b30 --- /dev/null +++ b/word-aligner/makefiles/makefile.model.f-e @@ -0,0 +1,14 @@ +all: output.f-e.aligned + +clean: + $(RM) output.f-e.a weights.cur.gz + +CDEC = $(SCRIPT_DIR)/../decoder/cdec +OPTIMIZE = $(SCRIPT_DIR)/../training/crf/mpi_online_optimize + +weights.cur.gz: ../grammars/wordpairs.f-e.features.gz + $(MPIRUN) $(OPTIMIZE) -a agenda.txt -1 -C 1.0 -t ../grammars/corpus.f-e --max_walltime 90 + +output.f-e.aligned: weights.cur.gz + $(CDEC) -c cdec.ini -w $< --lextrans_align_only -i ../grammars/corpus.f-e -a > $@ + |