fn
arange_op
→TensorImplPtrint arange_op(double start, double stop, double step, Dtype dt, Device device, bool requires_grad)Python wrappers
Public Python APIs implemented by this engine symbol.Create a 1-D tensor with values up to but not including stop.
Math
Output length is ; element equals .
Parameters
startdoubleFirst value of the progression.
stopdoubleExclusive upper bound.
stepdoubleConstant stride. Must be non-zero.
dtDtypeElement dtype (the values are cast at the end).
deviceDeviceTarget device.
requires_gradboolAttach a fresh autograd leaf when true.
Returns
TensorImplPtr1-D tensor of length .
Raises
LucidErrorIf
step == 0.