cuda.FloatTensor constructor with float32 array #1957
Description
FloatTensor(np.array([0,1,2], dtype=np.float64))
cuda.FloatTensor(np.array([0,1,2], dtype=np.float64))
FloatTensor(np.array([0,1,2], dtype=np.float32))work as expected, while
cuda.FloatTensor(np.array([0,1,2], dtype=np.float32))fails with
RuntimeError: tried to construct a tensor from a float sequence, but found an item of type numpy.float32 at index (0)
PyTorch version: 0.1.12_2