search

CorticalComputer / DXNN2

Topology and Parameter Evolving Universal Learning Network platform DXNN MK2
www.DXNNResearch.com
Apache License 2.0
151 stars 53 forks source link

Can't compile #2

Open julienmarie opened 10 years ago

julienmarie commented 10 years ago

Hello, circuit.erl fails on compilation ( on erlang 17 ) ( after a long series of warnings ). Any idea ?

CorticalComputer commented 10 years ago

I was working on a number of new elements before making that last upload, and have not uploaded the final version. I'll try to do so today.

circuit.erl is not needed if the more advanced features are not used (modular neural networks, preprocessors...). So you can skip it for now, same with the complex.erl, and whatever else does not compile, can be skipped with the system continuing to work just fine. The modules that are not compiling were those that I was actively modifying before I made that last upload.

On Fri, Jul 4, 2014 at 7:05 AM, julienmarie notifications@github.com wrote:

Hello, circuit.erl fails on compilation ( on erlang 17 ) ( after a long series of warnings ). Any idea ?

— Reply to this email directly or view it on GitHub https://github.com/CorticalComputer/DXNN2/issues/2.

TeemuPentinsaari commented 10 years ago

Hi guys! Any timetable for this fix?

CorticalComputer commented 10 years ago

The system works with the non-compiling modules due to those modules non-essential nature. What file do you need to be fixed for system's functionality? I can send those files to you directly. The system as is, is fully functional without the non-compiling files, only the experimental and un-needed/unused files are not working since they are... experimental, and undocumented either way, in which I am exploring modular and unsupervised learning. I can create a separate track where I simply cut out all the files that are not necessary for standard functionality, and in which I am exploring new features (for example, circuit has code for deep learning, convolutional networks, stacked denoising autoencoders, applications of the same to bioinformatics... They are not yet documented in DXNN2, thus need not be worried about, same for complex.erl...).

On Tue, Aug 5, 2014 at 9:17 PM, TeemuPentinsaari notifications@github.com wrote:

Hi guys! Any timetable for this fix?

— Reply to this email directly or view it on GitHub https://github.com/CorticalComputer/DXNN2/issues/2#issuecomment-51282931 .

julienmarie commented 10 years ago

Hi Gene,

Indeed it would be really useful to have a separate branch for a clean build :)

Thanks!

Julien

2014-08-11 12:32 GMT+08:00 Gene notifications@github.com:

The system works with the non-compiling modules due to those modules non-essential nature. What file do you need to be fixed for system's functionality? I can send those files to you directly. The system as is, is fully functional without the non-compiling files, only the experimental and un-needed/unused files are not working since they are... experimental, and undocumented either way, in which I am exploring modular and unsupervised learning. I can create a separate track where I simply cut out all the files that are not necessary for standard functionality, and in which I am exploring new features (for example, circuit has code for deep learning, convolutional networks, stacked denoising autoencoders, applications of the same to bioinformatics... They are not yet documented in DXNN2, thus need not be worried about, same for complex.erl...).

On Tue, Aug 5, 2014 at 9:17 PM, TeemuPentinsaari notifications@github.com

wrote:

Hi guys! Any timetable for this fix?

— Reply to this email directly or view it on GitHub < https://github.com/CorticalComputer/DXNN2/issues/2#issuecomment-51282931> .

— Reply to this email directly or view it on GitHub https://github.com/CorticalComputer/DXNN2/issues/2#issuecomment-51741216 .

Julien Marie

jm@producture.com skype : thewiseup +63 917 511 8258

Book an appointment with me : Face to face meeting : https://www.timetrade.com/book/BM8SN Phone call : https://www.timetrade.com/book/KVW52

-----BEGIN PGP PUBLIC KEY BLOCK-----

xsBNBFNmiRIBCAC77MQjKyaDaRo9bQhfGNXIDK3SldtI+A56EAY2xh8/WCrJ HYwhdwW/2aConIa5eL92WLf0W6FSmjk0rqSp1bkkDt8RW7JFiftCACZkSwHF Rg3zpUeMgFG3BBRpwCiOxSygDza61AmMAQ+8LuLZOETEFddT/DQNCBouZLYN KE9uCs+lYkrpc3YbPfx+kGUDj5rT2SYnHwt4r352OjT1U8ABQdxBigymmzgT rZh/4oyVXjMtBCRwAqwmVtzwoNoarAcgsm6qQgNwTdy5NtmKlgs2hqQtIkQK 3uL+e3zyWb1B8e6gUhpn5cooeJG1hvlJ65XUTHgHOCCz7Ue4q1rl/5kZABEB AAHNIEp1bGllbiBNYXJpZSA8am1AcHJvZHVjdHVyZS5jb20+wsByBBABCAAm BQJTZokZBgsJCAcDAgkQ7nj49Sxgl2gEFQgCCgMWAgECGwMCHgEAANulB/4u c+UeSeo8Wn8COAKSoMyWh2m5zBEdRdTlELVNcQIH3ssqu1M3TEpfBixVofp2 NPYKNF7FrQPl7z9+w0azlVJVIDF6DlGWjVlKm1nDTlVr4gW8J6mgRYivV3Tp HlBHIFmelOIcBpJAhUfc73H5a9BAsJ/a2Iekpc0TERsnarVuA3sqeTqPkwBY komURJWx0yxiYj91Be1PDRRi+DyGBKa8AeapXiCmCtpGAxksB43iwiiFnLev iZCWp3bKeBCLZy5IkONNU6mNzTcljrgHzRjs9P8OSiEIJ9T8tubHUkkA0LsR 9l7vmBk3aESNF8w1LmRTRN+gBkXsMSjzNilGj7zBzsBNBFNmiRkBCACOsST4 AuZffy8Ny7w1Qhvo1Okt9bBoh5zIJKE0tEo1L4YmDYALKnagVHNUwiAM26Bv ynK+pAvRopt62Wu+s1hQgHNuwnTMuuL0OW4Y8daOIN9pUAqnT/etskSwH24W niP445IEa09RyA1b2Mwhy8hEN/eCFHmnY36U5+qiiUAjN6Khsflzkq4I2Qy9 wwMVhNC8/J1QtmODcUekMgnGG0XcXNm22N1EU3llr5yG/oYD0dAKnIIv4rf8 3s6LU7ND/SOjmaLNtzwLuGawcCBPGfaBpB77ozJz0jJ3RaEwuY/Kqt4Zfp5T 0xLFgr+fXxSjHku0qU6PGT3iH+eHkWZXGG2pABEBAAHCwF8EGAEIABMFAlNm iRwJEO54+PUsYJdoAhsMAADWtwf9HJiym4Oi2zwWiDksmLc7699iVY1pvuic RlqwMw2bmFd+PUUFupU6CR8jOTcYPJm3NzZ9gwPM3yzjd/BReWmd4mfZJfvf EIYHzJoDRmnT5t8b9Wp6jllWtVAHbN9FTUW3uJODDKhffhYNbmAhxpotao+a XQJcSrcWE1Gl030ZQryOfkz2l/eUfYYSK38gsR7ksz22Gb1CKS4gmG8jgxQs vmDTSRagUYEXOdTaX5HKMdHrMw1ll1Z8bMb/y0N0kgDmIF0pAX2i+bk5UpnU 6mbeLCwo0OoIPYpC4GvGPBglull+BRJlwvf9Kpzlip5KHYSTnb19yyjbs1qC rpsqrnUA6g== =o6AX -----END PGP PUBLIC KEY BLOCK-----

ausrasul commented 7 years ago

Hi Gene, Thanks for your answer the other day.

I found out that if I exclude the circuit.erl and other non-compilable files, the framework works, but you can't use plasticity like neuromodulation and all others.

The reason is that you added two zeros in the neuron weight tuple, it was {Weight, PlasticityFunction} and now it is {Weight, 0, 0, PlasticityFunction}. See genotype.erl line 210.

The commit that introduced this change is:

How do I go around this? just remove the zeros? It is difficult to see if this work around will have other side effects.

Appreciate the help.

ausrasul commented 7 years ago

Now I can see that neuron.erl uses the new weight format: neuron.erl line 271 perturb_weightsP(Spread,MP,[{W,PDW,LP,LPs}|Weights],Acc)->

So it is the plasticity.erl that should be fixed... not sure how.