PrincetonUniversity / PsyNeuLink / 13536727089 / 1

85%

master: 85%

push

github

tests/AutodiffComposition: Explicitly copy tensors to CPU before converting to numpy (#3217)

Fixes: `TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.`
When the test_training_xor_with_batching test is run on a CUDA enabled
machine.

Fixes: beebd2a96
	("Add support for batching to AutoDiffCompostion in PyTorch mode. (#3204)")

Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>

9639 of 12698 branches covered (75.91%)

Branch coverage included in aggregate %.

33494 of 38612 relevant lines covered (86.75%)

0.87 hits per line