StringLookup layer can now take tf.SparseTensors as input. (keras-tea…

…m#19764)

keras/src/layers/preprocessing/string_lookup.py

@@ -382,7 +382,7 @@ def get_config(self):
         return {**base_config, **config}
 
     def call(self, inputs):
-        if isinstance(inputs, (tf.Tensor, tf.RaggedTensor)):
+        if isinstance(inputs, (tf.Tensor, tf.RaggedTensor, tf.SparseTensor)):
             tf_inputs = True
         else:
             tf_inputs = False

keras/src/layers/preprocessing/string_lookup_test.py

@@ -1,4 +1,5 @@
 import numpy as np
+import pytest
 from tensorflow import data as tf_data
 
 from keras.src import backend
@@ -38,6 +39,26 @@ def test_fixed_vocabulary(self):
         self.assertTrue(backend.is_tensor(output))
         self.assertAllClose(output, np.array([2, 3, 0]))
 
+    @pytest.mark.skipif(
+        not backend.backend() == "tensorflow", reason="Requires tf.SparseTensor"
+    )
+    def test_sparse_inputs(self):
+        import tensorflow as tf
+
+        layer = layers.StringLookup(
+            output_mode="int",
+            vocabulary=["a", "b", "c"],
+        )
+        input_data = tf.SparseTensor(
+            indices=[[0, 0], [1, 1], [2, 2]],
+            values=["b", "c", "d"],
+            dense_shape=(3, 3),
+        )
+        output = layer(input_data)
+        self.assertIsInstance(output, tf.SparseTensor)
+        self.assertAllClose(output, np.array([[2, 0, 0], [0, 3, 0], [0, 0, 0]]))
+        self.assertAllClose(output.values, np.array([2, 3, 0]))
+
     def test_set_vocabulary(self):
         layer = layers.StringLookup(
             output_mode="int",