Yes I was talking about the Spatial Pooler, sorry to mix it in this
conversation.
Great that you are working on that!

On Tue, Nov 10, 2015 at 2:03 PM Chetan Surpur <[email protected]> wrote:

> Gari,
>
> In your examples you use `SpatialPooler`, so I'm going to assume you're
> talking about that (instead of about the C++ vs Python Temporal Memory).
> Regarding the Spatial Pooler, I'm actually currently working on making the
> C++ version (exposed via SWIG bindings) and the Python version expose
> identical interfaces, so you can use them interchangeably [1]. This work is
> in progress.
>
> [1] https://github.com/numenta/nupic.core/issues/688
>
> - Chetan
>
>
> On Nov 10, 2015, at 12:49 PM, Garikoitz Lerma-Usabiaga <
> [email protected]> wrote:
>
> Hi Chetan,
> are the CPP and PY version exactly the same? I am encountering several
> differences between these two:
>
> from nupic.research.spatial_pooler import SpatialPooler
> from nupic.bindings.algorithms import SpatialPooler as CPPSpatialPooler
>
> I don't know if it is the expected behaviour and if it is documented
> somewhere or I have different versions of the corresponding files...
> For example: wrapAround only exists in the PY version:
>                TypeError: __init__() got an unexpected keyword argument
> 'wrapAround'
>
> I would like to use the fast CPP version meanwhile I am reading the python
> version, but if they are not the same it doesn't make sense. If I want to
> change something or see the internal state I run the PY version with with
> break points.
>
> Thanks!
> Gari
>
>
>
>
> On Tue, Nov 10, 2015 at 12:14 PM Chetan Surpur <[email protected]>
> wrote:
>
>> Wakan,
>>
>> This file is pretty self-contained, and is meant to be readable to
>> someone who can read python code. Let me know if you have any questions.
>>
>> - Chetan
>>
>> > On Nov 10, 2015, at 11:42 AM, Wakan Tanka <[email protected]> wrote:
>> >
>> > Thank you Chetan,
>> > I suppose this is the only documentation :) or am I wrong? I mean if I
>> will be able to figure out what is going on under the hood after
>> understanding temporal_memory.py or do I need to understand whole NuPIC
>> code.
>> >
>> > Thank you
>> >
>> > On 11/10/2015 07:01 PM, Chetan Surpur wrote:
>> >> Hi Wakan,
>> >>
>> >> The documented code for the Temporal Memory algorithm is in nupic [1].
>> >>
>> >> [1]
>> >>
>> https://github.com/numenta/nupic/blob/master/src/nupic/research/temporal_memory.py
>> >>
>> >> - Chetan
>> >>
>> >
>> >
>>
>>
>>
>

Reply via email to