reminisce commented on a change in pull request #14638: [MXNET-1382] Add the
index_array operator
URL: https://github.com/apache/incubator-mxnet/pull/14638#discussion_r283200718
##########
File path: tests/python/unittest/test_operator.py
##########
@@ -8054,6 +8054,68 @@ def test_image_normalize():
# check backward using finite difference
check_numeric_gradient(img_norm_sym, [data_in_4d], atol=0.001)
+@with_seed()
+def test_index_array_default():
+ for shape in [(10,), (7, 5, 29), (5, 7, 11, 13, 17, 19)]:
+ data = mx.symbol.Variable("data")
+ index_array = mx.sym.contrib.index_array(data)
+
+ input_array = np.ones(shape)
+ mgrid = np.mgrid[tuple(slice(0, x) for x in shape)]
+ expected = np.stack(mgrid, axis=-1)
+
+ check_symbolic_forward(index_array, [input_array], [expected])
+ check_symbolic_backward(index_array, [input_array],
[np.ones(expected.shape)], [np.zeros_like(input_array)])
+
+@with_seed()
+def test_index_array_default_zero_dim():
+ with mx.np_compat(active=True):
Review comment:
If the whole function is expected to be numpy compatible, you can use a
decorator `mx.use_np_compat` to wrap the whole function. It would be easier for
us to just delete the decorator in the next major release than putting
additional efforts of adjusting indentation.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
