Picks the best algorithm based on device, and scrambles seed into key and counter.
This op picks the best counter-based RNG algorithm based on device, and scrambles a shape-[2] seed into a key and a counter, both needed by the counter-based algorithm. The scrambling is opaque but approximately satisfies the property that different seed results in different key/counter pair (which will in turn result in different random numbers).
Constants
String | OP_NAME | The name of this op, as known by TensorFlow core engine |
Public Methods
Output<TInt32> |
alg()
The RNG algorithm (shape int32[]).
|
Output<?> |
counter()
Counter for the counter-based RNG algorithm.
|
static StatelessRandomGetKeyCounterAlg | |
Output<?> |
key()
Key for the counter-based RNG algorithm (shape uint64[1]).
|
Inherited Methods
Constants
public static final String OP_NAME
The name of this op, as known by TensorFlow core engine
Public Methods
public Output<?> counter ()
Counter for the counter-based RNG algorithm. Since counter size is algorithm-dependent, this output will be right-padded with zeros to reach shape uint64[2] (the current maximal counter size among algorithms).
public static StatelessRandomGetKeyCounterAlg create (Scope scope, Operand<? extends TNumber> seed)
Factory method to create a class wrapping a new StatelessRandomGetKeyCounterAlg operation.
Parameters
scope | current scope |
---|---|
seed | 2 seeds (shape [2]). |
Returns
- a new instance of StatelessRandomGetKeyCounterAlg