[Numpy-discussion] printable random seed ?

Robert Kern robert.kern@gmail....
Wed Sep 22 10:53:22 CDT 2010


On Wed, Sep 22, 2010 at 10:48,  <josef.pktd@gmail.com> wrote:
> On Wed, Sep 22, 2010 at 10:34 AM, Robert Kern <robert.kern@gmail.com> wrote:
>> On Wed, Sep 22, 2010 at 09:27,  <josef.pktd@gmail.com> wrote:
>>> I would like to generate random numbers based on a random seed, for
>>> example what numpy.random does if the seed is not specified. But I
>>> would also like to print out the initial state, so I can replicate the
>>> random numbers.
>>>
>>> Can I get a human readable or printable version of the initial state?
>>
>> [~]
>> |13> prng = np.random.RandomState()
>>
>> [~]
>> |14> prng2 = np.random.RandomState()
>>
>> [~]
>> |15> prng.randint(100, size=10)
>> array([74, 62, 56, 94, 86, 59, 69, 94, 42, 18])
>>
>> [~]
>> |16> prng2.randint(100, size=10)
>> array([21, 58, 34, 55,  9, 81, 45,  3, 93, 62])
>>
>> [~]
>> |17> prng2.set_state(prng.get_state())
>>
>> [~]
>> |18> prng.randint(100, size=10)
>> array([37, 57,  2,  0, 68,  9, 75, 88, 11,  7])
>>
>> [~]
>> |19> prng2.randint(100, size=10)
>> array([37, 57,  2,  0, 68,  9, 75, 88, 11,  7])
>>
>> [~]
>> |20> prng.get_state()
>>
>> ('MT19937',
>>  array([1368120112,  957462593, 2623310617, 4207155283,  446940397,
>>       3506388262, 4104366519,  371500243, 4029407620,  899392379,
>> ....
>>       1843090101, 2484333397, 4085469971,  306955884,   23307203,
>>       1640066622,   48186677,  637144011,  854838500], dtype=uint32),
>>  26,
>>  0,
>>  2.5933437794758841e-288)
>>
>>> Alternatively, what's a good way to  randomly generate an initial
>>> state?
>>
>> np.random.RandomState() will do the best thing on each platform.
>
> Thanks, I need to think about whether I want to save the full
> get_state() tuple (or cheat)

Save the whole thing.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless
enigma that is made terrible by our own mad attempt to interpret it as
though it had an underlying truth."
  -- Umberto Eco


More information about the NumPy-Discussion mailing list