From 1041f68be2fe66c632c23e06472b0748990326f4 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Danie=CC=88l=20de=20Kok?= <me@danieldk.eu>
Date: Fri, 2 Dec 2022 16:30:03 +0100
Subject: [PATCH] Use cross entropy from `thinc.legacy`

---
 spacy/pipeline/edit_tree_lemmatizer.py | 7 +++++--
 spacy/pipeline/morphologizer.pyx       | 5 +++--
 spacy/pipeline/senter.pyx              | 6 ++++--
 spacy/pipeline/tagger.pyx              | 7 ++++---
 4 files changed, 16 insertions(+), 9 deletions(-)

diff --git a/spacy/pipeline/edit_tree_lemmatizer.py b/spacy/pipeline/edit_tree_lemmatizer.py
index 9676e2194..0531d4ba5 100644
--- a/spacy/pipeline/edit_tree_lemmatizer.py
+++ b/spacy/pipeline/edit_tree_lemmatizer.py
@@ -5,8 +5,9 @@ from itertools import islice
 import numpy as np
 
 import srsly
-from thinc.api import Config, Model, SequenceCategoricalCrossentropy
+from thinc.api import Config, Model
 from thinc.types import ArrayXd, Floats2d, Ints1d
+from thinc.legacy import LegacySequenceCategoricalCrossentropy
 
 from ._edit_tree_internals.edit_trees import EditTrees
 from ._edit_tree_internals.schemas import validate_edit_tree
@@ -129,7 +130,9 @@ class EditTreeLemmatizer(TrainablePipe):
         self, examples: Iterable[Example], scores: List[Floats2d]
     ) -> Tuple[float, List[Floats2d]]:
         validate_examples(examples, "EditTreeLemmatizer.get_loss")
-        loss_func = SequenceCategoricalCrossentropy(normalize=False, missing_value=-1)
+        loss_func = LegacySequenceCategoricalCrossentropy(
+            normalize=False, missing_value=-1
+        )
 
         truths = []
         for eg in examples:
diff --git a/spacy/pipeline/morphologizer.pyx b/spacy/pipeline/morphologizer.pyx
index 782a1dabe..293add9e1 100644
--- a/spacy/pipeline/morphologizer.pyx
+++ b/spacy/pipeline/morphologizer.pyx
@@ -1,7 +1,8 @@
 # cython: infer_types=True, profile=True, binding=True
 from typing import Callable, Dict, Iterable, List, Optional, Union
 import srsly
-from thinc.api import SequenceCategoricalCrossentropy, Model, Config
+from thinc.api import Model, Config
+from thinc.legacy import LegacySequenceCategoricalCrossentropy
 from thinc.types import Floats2d, Ints1d
 from itertools import islice
 
@@ -290,7 +291,7 @@ class Morphologizer(Tagger):
         DOCS: https://spacy.io/api/morphologizer#get_loss
         """
         validate_examples(examples, "Morphologizer.get_loss")
-        loss_func = SequenceCategoricalCrossentropy(names=tuple(self.labels), normalize=False)
+        loss_func = LegacySequenceCategoricalCrossentropy(names=tuple(self.labels), normalize=False)
         truths = []
         for eg in examples:
             eg_truths = []
diff --git a/spacy/pipeline/senter.pyx b/spacy/pipeline/senter.pyx
index 93a7ee796..42feeb277 100644
--- a/spacy/pipeline/senter.pyx
+++ b/spacy/pipeline/senter.pyx
@@ -3,7 +3,9 @@ from typing import Dict, Iterable, Optional, Callable, List, Union
 from itertools import islice
 
 import srsly
-from thinc.api import Model, SequenceCategoricalCrossentropy, Config
+from thinc.api import Model, Config
+from thinc.legacy import LegacySequenceCategoricalCrossentropy
+
 from thinc.types import Floats2d, Ints1d
 
 from ..tokens.doc cimport Doc
@@ -161,7 +163,7 @@ class SentenceRecognizer(Tagger):
         """
         validate_examples(examples, "SentenceRecognizer.get_loss")
         labels = self.labels
-        loss_func = SequenceCategoricalCrossentropy(names=labels, normalize=False)
+        loss_func = LegacySequenceCategoricalCrossentropy(names=labels, normalize=False)
         truths = []
         for eg in examples:
             eg_truth = []
diff --git a/spacy/pipeline/tagger.pyx b/spacy/pipeline/tagger.pyx
index 3b4715ce5..e12f116af 100644
--- a/spacy/pipeline/tagger.pyx
+++ b/spacy/pipeline/tagger.pyx
@@ -2,7 +2,8 @@
 from typing import Callable, Dict, Iterable, List, Optional, Union
 import numpy
 import srsly
-from thinc.api import Model, set_dropout_rate, SequenceCategoricalCrossentropy, Config
+from thinc.api import Model, set_dropout_rate, Config
+from thinc.legacy import LegacySequenceCategoricalCrossentropy
 from thinc.types import Floats2d, Ints1d
 import warnings
 from itertools import islice
@@ -244,7 +245,7 @@ class Tagger(TrainablePipe):
 
         DOCS: https://spacy.io/api/tagger#rehearse
         """
-        loss_func = SequenceCategoricalCrossentropy()
+        loss_func = LegacySequenceCategoricalCrossentropy()
         if losses is None:
             losses = {}
         losses.setdefault(self.name, 0.0)
@@ -275,7 +276,7 @@ class Tagger(TrainablePipe):
         DOCS: https://spacy.io/api/tagger#get_loss
         """
         validate_examples(examples, "Tagger.get_loss")
-        loss_func = SequenceCategoricalCrossentropy(names=self.labels, normalize=False, neg_prefix=self.cfg["neg_prefix"])
+        loss_func = LegacySequenceCategoricalCrossentropy(names=self.labels, normalize=False, neg_prefix=self.cfg["neg_prefix"])
         # Convert empty tag "" to missing value None so that both misaligned
         # tokens and tokens with missing annotation have the default missing
         # value None.