Merge pull request #36316 from faustomorales:fix/#35799-ctc-decode

PiperOrigin-RevId: 313619800
Change-Id: I4843b83f42a5425d722d9b509d241941745fb5ff

tensorflow/python/keras/backend.py

@@ -5948,10 +5948,13 @@ def ctc_decode(y_pred, input_length, greedy=True, beam_width=100, top_paths=1):
               contains the decoded sequence.
               If `false`, returns the `top_paths` most probable
               decoded sequences.
+              Each decoded sequence has shape (samples, time_steps).
               Important: blank labels are returned as `-1`.
           Tensor `(top_paths, )` that contains
               the log probability of each decoded sequence.
   """
+  input_shape = shape(y_pred)
+  samples, steps = input_shape[0], input_shape[1]
   y_pred = math_ops.log(array_ops.transpose(y_pred, perm=[1, 0, 2]) + epsilon())
   input_length = math_ops.cast(input_length, dtypes_module.int32)
 
@@ -5966,7 +5969,7 @@ def ctc_decode(y_pred, input_length, greedy=True, beam_width=100, top_paths=1):
         top_paths=top_paths)
   decoded_dense = [
       sparse_ops.sparse_to_dense(
-          st.indices, st.dense_shape, st.values, default_value=-1)
+          st.indices, (samples, steps), st.values, default_value=-1)
       for st in decoded
   ]
   return (decoded_dense, log_prob)

tensorflow/python/keras/backend_test.py

@@ -1762,7 +1762,10 @@ class TestCTC(test.TestCase):
         -3.777835    # output beam 1
     ], np.float32)[np.newaxis, :]
 
-    decode_truth = [np.array([1, 0]), np.array([0, 1, 0])]
+    decode_truth = [
+        np.array([1, 0, -1, -1, -1, -1, -1]),
+        np.array([0, 1, 0, -1, -1, -1, -1])
+    ]
     beam_width = 2
     top_paths = 2