[numpy] Fix users of NumPy APIs that are removed in NumPy 2.0.

This change migrates users of APIs removed in NumPy 2.0 to their recommended replacements (https://numpy.org/devdocs/numpy_2_0_migration_guide.html).

PiperOrigin-RevId: 656419848

tensorflow_probability/python/bijectors/bijector_test_util.py

@@ -89,8 +89,12 @@ def assert_scalar_congruency(bijector,
   ten_x_pts = np.linspace(lower_x, upper_x, num=10).astype(np.float32)
   if bijector.dtype is not None:
     ten_x_pts = ten_x_pts.astype(dtype_util.as_numpy_dtype(bijector.dtype))
-    lower_x = np.cast[dtype_util.as_numpy_dtype(bijector.dtype)](lower_x)
-    upper_x = np.cast[dtype_util.as_numpy_dtype(bijector.dtype)](upper_x)
+    lower_x = np.asarray(
+        lower_x, dtype=dtype_util.as_numpy_dtype(bijector.dtype)
+    )
+    upper_x = np.asarray(
+        upper_x, dtype=dtype_util.as_numpy_dtype(bijector.dtype)
+    )
   forward_on_10_pts = bijector.forward(ten_x_pts)
 
   # Set the lower/upper limits in the range of the bijector.

tensorflow_probability/python/distributions/lambertw_f_test.py

@@ -205,7 +205,7 @@ def dist_lambda(t):
     model.compile(optimizer=optimizer, loss=negloglik)
 
     model.fit(x, y, epochs=1, verbose=True, batch_size=32, validation_split=0.2)
-    self.assertGreater(model.history.history["val_loss"][0], -np.Inf)
+    self.assertGreater(model.history.history["val_loss"][0], -np.inf)
 
 
 if __name__ == "__main__":

tensorflow_probability/python/internal/dtype_util.py

@@ -231,6 +231,15 @@ def _unify_dtype(current, new):
       lambda dt, h: base_dtype(h if dt is None else dt), dtype, dtype_hint)
 
 
+def _issctype(x):
+  if not isinstance(x, (type, np.dtype)):
+    return False
+  try:
+    return np.dtype(x) != np.object_
+  except:  # pylint: disable=bare-except
+    return False
+
+
 def convert_to_dtype(tensor_or_dtype, dtype=None, dtype_hint=None):
   """Get a dtype from a list/tensor/dtype using convert_to_tensor semantics."""
   if tensor_or_dtype is None:
@@ -244,7 +253,7 @@ def convert_to_dtype(tensor_or_dtype, dtype=None, dtype_hint=None):
   # Numpy dtypes defer to dtype/dtype_hint
   elif isinstance(tensor_or_dtype, np.ndarray):
     dt = base_dtype(dtype or dtype_hint or tensor_or_dtype.dtype)
-  elif np.issctype(tensor_or_dtype):
+  elif _issctype(tensor_or_dtype):
     dt = base_dtype(dtype or dtype_hint or tensor_or_dtype)
   else:
     # If this is a Python object, call `convert_to_tensor` and grab the dtype.

tensorflow_probability/python/internal/dtype_util_test.py

@@ -179,7 +179,9 @@ def testMax(self, dtype, expected_maxval):
       disable_numpy=True,
       reason='`convert_to_tensor` respects array dtypes in numpy backend.')
   def testConvertToDtype(self, tensor_or_dtype, dtype, dtype_hint):
-    if np.issctype(tensor_or_dtype):
+    if isinstance(tensor_or_dtype, np.generic) or hasattr(
+        tensor_or_dtype, 'dtype'
+    ):
       example_tensor = np.zeros([], tensor_or_dtype)
     elif isinstance(tensor_or_dtype, tf.DType):
       example_tensor = tf.zeros([], tensor_or_dtype)
@@ -203,7 +205,7 @@ def testConvertToDtype(self, tensor_or_dtype, dtype, dtype_hint):
       disable_jax=True,
       reason='`convert_to_tensor` only raises in TF mode.')
   def testConvertToDTypeRaises(self, tensor_or_dtype, dtype, dtype_hint):
-    if np.issctype(tensor_or_dtype):
+    if isinstance(tensor_or_dtype, np.generic):
       example_tensor = np.zeros([], tensor_or_dtype)
     elif isinstance(tensor_or_dtype, tf.DType):
       example_tensor = tf.zeros([], tensor_or_dtype)