diff --git a/bin/parser/train.py b/bin/parser/train.py
index 267b26275..489d9259c 100755
--- a/bin/parser/train.py
+++ b/bin/parser/train.py
@@ -165,17 +165,29 @@ def evaluate(Language, gold_tuples, model_dir, gold_preproc=False, verbose=False
     return scorer
 
 
-def write_parses(Language, dev_loc, model_dir, out_loc):
-    nlp = Language()
-    gold_tuples = read_docparse_file(dev_loc)
+def write_parses(Language, dev_loc, model_dir, out_loc, beam_width=None):
+    nlp = Language(data_dir=model_dir)
+    if beam_width is not None:
+        nlp.parser.cfg.beam_width = beam_width
+    gold_tuples = read_json_file(dev_loc)
     scorer = Scorer()
     out_file = codecs.open(out_loc, 'w', 'utf8')
-    for raw_text, segmented_text, annot_tuples in gold_tuples:
-        tokens = nlp(raw_text)
-        for t in tokens:
-            out_file.write(
-                '%s\t%s\t%s\t%s\n' % (t.orth_, t.tag_, t.head.orth_, t.dep_)
-            )
+    for raw_text, sents in gold_tuples:
+        sents = _merge_sents(sents)
+        for annot_tuples, brackets in sents:
+            if raw_text is None:
+                tokens = nlp.tokenizer.tokens_from_list(annot_tuples[1])
+                nlp.tagger(tokens)
+                nlp.entity(tokens)
+                nlp.parser(tokens)
+            else:
+                tokens = nlp(raw_text, merge_mwes=False)
+            gold = GoldParse(tokens, annot_tuples)
+            scorer.score(tokens, gold, verbose=False)
+            for t in tokens:
+                out_file.write(
+                    '%s\t%s\t%s\t%s\n' % (t.orth_, t.tag_, t.head.orth_, t.dep_)
+                )
     return scorer
 
 
@@ -204,7 +216,7 @@ def main(train_loc, dev_loc, model_dir, n_sents=0, n_iter=15, out_loc="", verbos
               corruption_level=corruption_level, n_iter=n_iter,
               beam_width=beam_width)
     if out_loc:
-        write_parses(English, dev_loc, model_dir, out_loc)
+        write_parses(English, dev_loc, model_dir, out_loc, beam_width=beam_width)
     scorer = evaluate(English, list(read_json_file(dev_loc)),
                       model_dir, gold_preproc=gold_preproc, verbose=verbose,
                       beam_width=beam_width)