correct some description of annotations and spelling of class name

CHANGELOG.md

@@ -74,7 +74,7 @@ and less prone to user error.
 
 ### Changed
 
--   Fixed the bug in `tfrs.layers.loss.SamplingProbablityCorrection` that logits
+-   Fixed the bug in `tfrs.layers.loss.SamplingProbabilityCorrection` that logits
     should subtract the log of item probability.
 -   `tfrs.experimental.optimizers.CompositeOptimizer`: an optimizer that
     composes multiple individual optimizers which can be applied to different

tensorflow_recommenders/layers/feature_interaction/dot_interaction.py

@@ -26,7 +26,7 @@ class DotInteraction(tf.keras.layers.Layer):
   section 2.1.3. Sparse activations and dense activations are combined.
   Dot interaction is applied to a batch of input Tensors [e1,...,e_k] of the
   same dimension and the output is a batch of Tensors with all distinct pairwise
-  dot products of the form dot(e_i, e_j) for i <= j if self self_interaction is
+  dot products of the form dot(e_i, e_j) for i <= j if self_interaction is
   True, otherwise dot(e_i, e_j) i < j.
 
   Attributes:

tensorflow_recommenders/layers/loss.py

@@ -147,7 +147,7 @@ def call(self, labels: tf.Tensor, logits: tf.Tensor,
     return logits + duplicate * MIN_FLOAT
 
 
-class SamplingProbablityCorrection(tf.keras.layers.Layer):
+class SamplingProbabilityCorrection(tf.keras.layers.Layer):
   """Sampling probability correction."""
 
   def __call__(self, logits: tf.Tensor,
@@ -156,3 +156,4 @@ def __call__(self, logits: tf.Tensor,
 
     return logits - tf.math.log(
         tf.clip_by_value(candidate_sampling_probability, 1e-6, 1.))
+

tensorflow_recommenders/layers/loss_test.py

@@ -112,15 +112,15 @@ def test_sampling_probability_correction(self, random_seed):
     logits = rng.uniform(size=shape).astype(np.float32)
     probs = rng.uniform(size=shape[0]).astype(np.float32)
 
-    corrected_logits = loss.SamplingProbablityCorrection()(logits, probs)
+    corrected_logits = loss.SamplingProbabilityCorrection()(logits, probs)
     corrected_logits = corrected_logits.numpy()
 
     np.testing.assert_array_less(logits, corrected_logits)
 
     # set some of the probabilities to 0
     probs_with_zeros = probs * rng.choice([0., 1.], size=probs.shape)
 
-    corrected_logits_with_zeros = loss.SamplingProbablityCorrection()(
+    corrected_logits_with_zeros = loss.SamplingProbabilityCorrection()(
         logits, probs_with_zeros)
     corrected_logits_with_zeros = corrected_logits_with_zeros.numpy()
 

tensorflow_recommenders/tasks/retrieval.py

@@ -122,7 +122,7 @@ def call(self,
     Args:
       query_embeddings: [num_queries, embedding_dim] tensor of query
         representations.
-      candidate_embeddings: [num_queries, embedding_dim] tensor of candidate
+      candidate_embeddings: [num_candidates, embedding_dim] tensor of candidate
         representations.
       sample_weight: [num_queries] tensor of sample weights.
       candidate_sampling_probability: Optional tensor of candidate sampling
@@ -150,7 +150,7 @@ def call(self,
       scores = scores / self._temperature
 
     if candidate_sampling_probability is not None:
-      scores = layers.loss.SamplingProbablityCorrection()(
+      scores = layers.loss.SamplingProbabilityCorrection()(
           scores, candidate_sampling_probability)
 
     if self._remove_accidental_hits: