summaryrefslogtreecommitdiff
path: root/training
diff options
context:
space:
mode:
authorgraehl <graehl@ec762483-ff6d-05da-a07a-a48fb63a330f>2010-07-01 19:22:06 +0000
committergraehl <graehl@ec762483-ff6d-05da-a07a-a48fb63a330f>2010-07-01 19:22:06 +0000
commit75153dfed4623ac3e5da76675bb2475dd631d22b (patch)
tree60d3135f9108046cd65ea2ca5546278f50c4dba7 /training
parent043986db93d7f49ab1b1017a2322767d56384f7f (diff)
collapse_weights --unnormalized
git-svn-id: https://ws10smt.googlecode.com/svn/trunk@91 ec762483-ff6d-05da-a07a-a48fb63a330f
Diffstat (limited to 'training')
-rw-r--r--training/collapse_weights.cc14
1 files changed, 12 insertions, 2 deletions
diff --git a/training/collapse_weights.cc b/training/collapse_weights.cc
index 5e0f3f72..4fb742fb 100644
--- a/training/collapse_weights.cc
+++ b/training/collapse_weights.cc
@@ -1,3 +1,10 @@
+char const* NOTES =
+ "ZF_and_E means unnormalized scaled features.\n"
+ "For grammars with one nonterminal: F_and_E is joint,\n"
+ "F_given_E and E_given_F are conditional.\n"
+ "TODO: group rules by root nonterminal and then normalize.\n";
+
+
#include <iostream>
#include <fstream>
#include <tr1/unordered_map>
@@ -20,7 +27,9 @@ void InitCommandLine(int argc, char** argv, po::variables_map* conf) {
po::options_description opts("Configuration options");
opts.add_options()
("grammar,g", po::value<string>(), "Grammar file")
- ("weights,w", po::value<string>(), "Weights file");
+ ("weights,w", po::value<string>(), "Weights file")
+ ("unnormalized,u", "Always include ZF_and_E unnormalized score (default: only if sum was >1)")
+ ;
po::options_description clo("Command line options");
clo.add_options()
("config,c", po::value<string>(), "Configuration file")
@@ -40,6 +49,7 @@ void InitCommandLine(int argc, char** argv, po::variables_map* conf) {
if (conf->count("help") || !conf->count("grammar") || !conf->count("weights")) {
cerr << dcmdline_options << endl;
+ cerr << NOTES << endl;
exit(1);
}
}
@@ -91,7 +101,7 @@ int main(int argc, char** argv) {
tr.scores_.clear();
cout << tr.AsString() << " ||| F_and_E=" << lp - log(tot);
- if (!normalized) {
+ if (!normalized || conf.count("unnormalized")) {
cout << ";ZF_and_E=" << lp;
}
cout << ";F_given_E=" << lp - log(e_tots[tr.e_])