Support floret for PretrainVectors (#12435)

* Support floret for PretrainVectors

* Format

spacy/errors.py

@@ -549,8 +549,8 @@ class Errors(metaclass=ErrorsWithCodes):
             "during training, make sure to include it in 'annotating components'")
 
     # New errors added in v3.x
-    E850 = ("The PretrainVectors objective currently only supports default "
+    E850 = ("The PretrainVectors objective currently only supports default or "
-            "vectors, not {mode} vectors.")
+            "floret vectors, not {mode} vectors.")
     E851 = ("The 'textcat' component labels should only have values of 0 or 1, "
             "but found value of '{val}'.")
     E852 = ("The tar file pulled from the remote attempted an unsafe path "

spacy/ml/models/multi_task.py

@@ -1,5 +1,5 @@
 from typing import Any, Optional, Iterable, Tuple, List, Callable, TYPE_CHECKING, cast
-from thinc.types import Floats2d
+from thinc.types import Floats2d, Ints1d
 from thinc.api import chain, Maxout, LayerNorm, Softmax, Linear, zero_init, Model
 from thinc.api import MultiSoftmax, list2array
 from thinc.api import to_categorical, CosineDistance, L2Distance
@@ -7,7 +7,7 @@ from thinc.loss import Loss
 
 from ...util import registry, OOV_RANK
 from ...errors import Errors
-from ...attrs import ID
+from ...attrs import ID, ORTH
 from ...vectors import Mode as VectorsMode
 
 import numpy
@@ -24,8 +24,6 @@ def create_pretrain_vectors(
     maxout_pieces: int, hidden_size: int, loss: str
 ) -> Callable[["Vocab", Model], Model]:
     def create_vectors_objective(vocab: "Vocab", tok2vec: Model) -> Model:
-        if vocab.vectors.mode != VectorsMode.default:
-            raise ValueError(Errors.E850.format(mode=vocab.vectors.mode))
         if vocab.vectors.shape[1] == 0:
             raise ValueError(Errors.E875)
         model = build_cloze_multi_task_model(
@@ -70,14 +68,23 @@ def get_vectors_loss(ops, docs, prediction, distance):
     """Compute a loss based on a distance between the documents' vectors and
     the prediction.
     """
-    # The simplest way to implement this would be to vstack the
+    vocab = docs[0].vocab
-    # token.vector values, but that's a bit inefficient, especially on GPU.
+    if vocab.vectors.mode == VectorsMode.default:
-    # Instead we fetch the index into the vectors table for each of our tokens,
+        # The simplest way to implement this would be to vstack the
-    # and look them up all at once. This prevents data copying.
+        # token.vector values, but that's a bit inefficient, especially on GPU.
-    ids = ops.flatten([doc.to_array(ID).ravel() for doc in docs])
+        # Instead we fetch the index into the vectors table for each of our
-    target = docs[0].vocab.vectors.data[ids]
+        # tokens, and look them up all at once. This prevents data copying.
-    target[ids == OOV_RANK] = 0
+        ids = ops.flatten([doc.to_array(ID).ravel() for doc in docs])
-    d_target, loss = distance(prediction, target)
+        target = docs[0].vocab.vectors.data[ids]
+        target[ids == OOV_RANK] = 0
+        d_target, loss = distance(prediction, target)
+    elif vocab.vectors.mode == VectorsMode.floret:
+        keys = ops.flatten([cast(Ints1d, doc.to_array(ORTH)) for doc in docs])
+        target = vocab.vectors.get_batch(keys)
+        target = ops.as_contig(target)
+        d_target, loss = distance(prediction, target)
+    else:
+        raise ValueError(Errors.E850.format(mode=vocab.vectors.mode))
     return loss, d_target
 
 

spacy/tests/training/test_pretraining.py

@@ -359,19 +359,15 @@ def test_pretrain_default_vectors():
     nlp.vocab.vectors = Vectors(shape=(10, 10))
     create_pretrain_vectors(1, 1, "cosine")(nlp.vocab, nlp.get_pipe("tok2vec").model)
 
+    # floret vectors are supported
+    nlp.vocab.vectors = Vectors(
+        data=get_current_ops().xp.zeros((10, 10)), mode="floret", hash_count=1
+    )
+    create_pretrain_vectors(1, 1, "cosine")(nlp.vocab, nlp.get_pipe("tok2vec").model)
+
     # error for no vectors
     with pytest.raises(ValueError, match="E875"):
         nlp.vocab.vectors = Vectors()
         create_pretrain_vectors(1, 1, "cosine")(
             nlp.vocab, nlp.get_pipe("tok2vec").model
         )
-
-    # error for floret vectors
-    with pytest.raises(ValueError, match="E850"):
-        ops = get_current_ops()
-        nlp.vocab.vectors = Vectors(
-            data=ops.xp.zeros((10, 10)), mode="floret", hash_count=1
-        )
-        create_pretrain_vectors(1, 1, "cosine")(
-            nlp.vocab, nlp.get_pipe("tok2vec").model
-        )