MCPcopy Create free account
hub / github.com/apache/singa / prepare_inputs_targets_for_rnn_test

Function prepare_inputs_targets_for_rnn_test

test/python/test_operation.py:61–84  ·  view source on GitHub ↗
(dev)

Source from the content-addressed store, hash-verified

59
60
61def prepare_inputs_targets_for_rnn_test(dev):
62 x_0 = np.random.random((2, 3)).astype(np.float32)
63 x_1 = np.random.random((2, 3)).astype(np.float32)
64 x_2 = np.random.random((2, 3)).astype(np.float32)
65
66 h_0 = np.zeros((2, 2)).astype(np.float32)
67
68 t_0 = np.random.random((2, 2)).astype(np.float32)
69 t_1 = np.random.random((2, 2)).astype(np.float32)
70 t_2 = np.random.random((2, 2)).astype(np.float32)
71
72 x0 = tensor.Tensor(device=dev, data=x_0)
73 x1 = tensor.Tensor(device=dev, data=x_1)
74 x2 = tensor.Tensor(device=dev, data=x_2)
75
76 h0 = tensor.Tensor(device=dev, data=h_0)
77
78 t0 = tensor.Tensor(device=dev, data=t_0)
79 t1 = tensor.Tensor(device=dev, data=t_1)
80 t2 = tensor.Tensor(device=dev, data=t_2)
81
82 inputs = [x0, x1, x2]
83 targets = [t0, t1, t2]
84 return inputs, targets, h0
85
86
87class TestPythonOperation(unittest.TestCase):

Calls 1

TensorMethod · 0.80

Tested by

no test coverage detected