anirudh2290 commented on a change in pull request #13362: Add NHWC layout
support to Pooling (cuDNN only)
URL: https://github.com/apache/incubator-mxnet/pull/13362#discussion_r239662668
##########
File path: tests/python/gpu/test_operator_gpu.py
##########
@@ -608,6 +608,52 @@ def test_convolution_versions():
@with_seed()
+def test_pooling_with_convention():
+ # While the float32 and float64 output is reliably consistent, float16
departs occasionally.
+ # We compare cpu and gpu results only within a given precision.
+ for data_type in [np.float64, np.float32, np.float16]:
+ ctx_list = [{'ctx': mx.gpu(0), 'pool_data': (2, 2, 10, 10),
'type_dict': {'pool_data': data_type}},
+ {'ctx': mx.cpu(0), 'pool_data': (2, 2, 10, 10),
'type_dict': {'pool_data': data_type}}]
+ sym = mx.sym.Pooling(kernel=(3,3), pool_type='max',
pooling_convention='valid', name='pool')
+ check_consistency(sym, ctx_list)
+
+ sym = mx.sym.Pooling(kernel=(3,3), pool_type='max',
pooling_convention='full', name='pool')
+ check_consistency(sym, ctx_list)
+
+ sym = mx.sym.Pooling(kernel=(300,300), pool_type='max',
global_pool=True, name='pool')
+ check_consistency(sym, ctx_list)
+
+
+@with_seed()
+@assert_raises_cudnn_not_satisfied(min_version='7.0.1')
Review comment:
This should be fine. The decorator checks if an exception is raised when
cudnn version is less than 7.0.1 and you run the test. For cudnn versions
greater than 7.0.1 it should just work like other tests.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on 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
