Fixed cases where tf.TensorShape was constructed with float dimensions

This is a prerequisite for making TensorShape and Dimension more strict
about the types of their arguments.

PiperOrigin-RevId: 274700832

tensorflow/python/distribute/all_reduce_test.py

@@ -123,12 +123,12 @@ class AllReduceTest(test_util.TensorFlowTestCase):
     for otl in output_tensors:
       self.assertEqual(len(otl), num_chunks)
       for ot in otl:
-        self.assertEqual(ot.shape, [tlen/num_chunks])
+        self.assertEqual(ot.shape, [tlen//num_chunks])
 
   def _buildInitialVars(self, shape, dev_list):
     values = []
     num_devices = len(dev_list)
-    dim = np.prod(shape) if shape else 1
+    dim = np.prod(shape, dtype=int) if shape else 1
     for d in range(0, num_devices):
       with ops.device(dev_list[d]):
         npt = np.zeros(shape).astype(np.float32)

tensorflow/python/tpu/tpu_sharding.py

@@ -184,7 +184,7 @@ class ShardingPolicy(object):
       raise ValueError("shape %s cannot be sharded %d ways along dimension %d" %
                        (shape.as_list(), self._number_of_shards,
                         self._shard_dimension))
-    dims[self._shard_dimension] /= self._number_of_shards
+    dims[self._shard_dimension] //= self._number_of_shards
     return tensor_shape.as_shape(dims)
 
   def _unshard_shape(self, shape):

tensorflow/python/training/input.py

@@ -398,7 +398,7 @@ class _SparseMetaData(object):
     """
     self._sparse = sparse
     self._map_op = map_op
-    self._rank = tensor_shape.Dimension(rank)
+    self._rank = tensor_shape.as_dimension(rank)
 
   def __eq__(self, other):
     if self.sparse != other.sparse: