skip alignment stage if the consolidated file exists
diff --git a/scripts/training/pipeline.pl b/scripts/training/pipeline.pl
index f7fa346..4c6380c 100755
--- a/scripts/training/pipeline.pl
+++ b/scripts/training/pipeline.pl
@@ -766,7 +766,13 @@
# --no-tokenization flag
}
-# skip this step if an alignment was provided
+# Use an existing alignment file if it's present, short-circuits
+# rebuilding the alignments...
+if (-s "alignments/training.align") {
+ $ALIGNMENT = "alignments/training.align";
+}
+
+# skip this step if an alignment was provided or it already exists
if (! defined $ALIGNMENT) {
# We process the data in chunks which by default are 1,000,000 sentence pairs. So first split up
