From 38e5a750145ae58e8e8a1997c815d5fc4f55b90d Mon Sep 17 00:00:00 2001
From: vinit <vinit.ravishankar@gmail.com>
Date: Fri, 17 Feb 2023 00:53:17 +0530
Subject: [PATCH] use True/False instead of floats

---
 spacy/pipeline/tagger.pyx           |  8 ++++----
 spacy/tests/pipeline/test_tagger.py | 11 ++++++++---
 2 files changed, 12 insertions(+), 7 deletions(-)

diff --git a/spacy/pipeline/tagger.pyx b/spacy/pipeline/tagger.pyx
index a877062fa..ae59d0983 100644
--- a/spacy/pipeline/tagger.pyx
+++ b/spacy/pipeline/tagger.pyx
@@ -45,7 +45,7 @@ DEFAULT_TAGGER_MODEL = Config().from_str(default_model_config)["model"]
 @Language.factory(
     "tagger",
     assigns=["token.tag"],
-    default_config={"model": DEFAULT_TAGGER_MODEL, "overwrite": False, "scorer": {"@scorers": "spacy.tagger_scorer.v1"}, "neg_prefix": "!", "label_smoothing": 0.0},
+    default_config={"model": DEFAULT_TAGGER_MODEL, "overwrite": False, "scorer": {"@scorers": "spacy.tagger_scorer.v1"}, "neg_prefix": "!", "label_smoothing": False},
     default_score_weights={"tag_acc": 1.0},
 )
 def make_tagger(
@@ -55,7 +55,7 @@ def make_tagger(
     overwrite: bool,
     scorer: Optional[Callable],
     neg_prefix: str,
-    label_smoothing: float,
+    label_smoothing: bool,
 ):
     """Construct a part-of-speech tagger component.
 
@@ -90,7 +90,7 @@ class Tagger(TrainablePipe):
         overwrite=BACKWARD_OVERWRITE,
         scorer=tagger_score,
         neg_prefix="!",
-        label_smoothing=0.0,
+        label_smoothing=False,
     ):
         """Initialize a part-of-speech tagger.
 
@@ -258,7 +258,7 @@ class Tagger(TrainablePipe):
         DOCS: https://spacy.io/api/tagger#get_loss
         """
         validate_examples(examples, "Tagger.get_loss")
-        # label_smoothing = 0.1 if self.cfg["label_smoothing"] else 0.0
+        self.cfg["label_smoothing"] = 0.05 if self.cfg["label_smoothing"] else 0.0
         loss_func = SequenceCategoricalCrossentropy(names=self.labels, normalize=False, neg_prefix=self.cfg["neg_prefix"], label_smoothing=self.cfg["label_smoothing"])
         # Convert empty tag "" to missing value None so that both misaligned
         # tokens and tokens with missing annotation have the default missing
diff --git a/spacy/tests/pipeline/test_tagger.py b/spacy/tests/pipeline/test_tagger.py
index 04284988f..32cdd196b 100644
--- a/spacy/tests/pipeline/test_tagger.py
+++ b/spacy/tests/pipeline/test_tagger.py
@@ -68,17 +68,22 @@ PARTIAL_DATA = [
 
 
 def test_label_smoothing():
+    util.fix_random_seed()
     nlp = Language()
-    tagger = nlp.add_pipe("tagger", config=dict(label_smoothing=True))
+    tagger_no_ls = nlp.add_pipe("tagger", "no_label_smoothing", config=dict(label_smoothing=False))
+    tagger_ls = nlp.add_pipe("tagger", "label_smoothing", config=dict(label_smoothing=True))
     train_examples = []
+    losses = {}
     for tag in TAGS:
-        tagger.add_label(tag)
+        tagger_no_ls.add_label(tag)
+        tagger_ls.add_label(tag)
     for t in TRAIN_DATA:
         train_examples.append(Example.from_dict(nlp.make_doc(t[0]), t[1]))
     optimizer = nlp.initialize(get_examples=lambda: train_examples)
-    for i in range(1):
+    for i in range(5):
         losses = {}
         nlp.update(train_examples, sgd=optimizer, losses=losses)
+    assert losses == {'no_label_smoothing': 1.4892945885658264, 'label_smoothing': 1.1432453989982605}
 
 
 def test_no_label():