* Update conll_train script

honnibal · honnibal · commit 77f2b218f9fb · 2016-02-24T18:19:38.000+01:00
diff --git a/bin/tagger/conll_train.py b/bin/tagger/conll_train.py
@@ -87,15 +87,15 @@ def _parse_line(line):
 
         
 def score_model(nlp, gold_tuples, verbose=False):
-    scorer = Scorer()
+    correct = 0.0
+    total = 0.0
     for words, gold_tags in gold_tuples:
         tokens = nlp.tokenizer.tokens_from_list(words)
         nlp.tagger(tokens)
         for token, gold in zip(tokens, gold_tags):
-            scorer.tags.tp += token.tag_ == gold
-            scorer.tags.fp += token.tag_ != gold
-            scorer.tags.fn += token.tag_ != gold
-    return scorer.tags_acc
+            correct += token.tag_ == gold
+            total += 1
+    return (correct / total) * 100
 
 
 def train(Language, train_sents, dev_sents, model_dir, n_iter=15, seed=0,
@@ -116,8 +116,6 @@ def train(Language, train_sents, dev_sents, model_dir, n_iter=15, seed=0,
     random.shuffle(train_sents)
     heldout_sents = train_sents[:int(nr_train * 0.1)] 
     train_sents = train_sents[len(heldout_sents):]
-    #train_sents = train_sents[:500]
-    #assert len(heldout_sents) < len(train_sents)
     prev_score = 0.0
     variance = 0.001
     last_good_learn_rate = nlp.tagger.model.eta
@@ -130,15 +128,15 @@ def train(Language, train_sents, dev_sents, model_dir, n_iter=15, seed=0,
         acc += nlp.tagger.train(tokens, gold_tags)
         total += len(tokens)
         n += 1
-        if n and n % 10000 == 0:
+        if n and n % 20000 == 0:
             dev_score = score_model(nlp, heldout_sents)
             eval_score = score_model(nlp, dev_sents)
-            if dev_score > prev_score:
+            if dev_score >= prev_score:
                 nlp.tagger.model.keep_update()
                 prev_score = dev_score
                 variance = 0.001
                 last_good_learn_rate = nlp.tagger.model.eta
-                nlp.tagger.model.eta *= 1.05
+                nlp.tagger.model.eta *= 1.01
                 print('%d:\t%.3f\t%.3f\t%.3f\t%.4f' % (n, acc/total, dev_score, eval_score, nlp.tagger.model.eta))
             else:
                 nlp.tagger.model.backtrack()
