Commit Graph

13 Commits

Author SHA1 Message Date
xdrm-brackets dee73f521a Optimization with @fitness, only keep fitnesses greater than @this->maxFit (if anti-regression-option enabled) + no_file anymore 2016-10-29 16:35:49 +02:00
xdrm-brackets 6826434e07 minmod 2016-10-29 15:47:32 +02:00
xdrm-brackets 1c7f4e6566 Corrections [Genome+NeuralNetwork+NeuralNetworkCore] + optimization 2016-10-28 15:35:55 +02:00
xdrm-brackets 7aefda0bf0 Method to get trained genome + anti-regression management + some stuff to do (output neurons issue) + manage `mutThr` decreasing + ... 2016-10-28 13:03:34 +02:00
xdrm-brackets 10f2a3951d Add .gitignore 2016-10-27 21:28:09 +02:00
xdrm-brackets da50b03202 Gestion de quasi-tout -> FONCTIONNELLLLLLL 2016-10-27 18:34:28 +02:00
xdrm-brackets e16048b569 Management of genome's training with callback + fitness user-defined + generation&genomes management to do in NeuralNetworkCore::learn() 2016-10-27 14:47:08 +02:00
xdrm-brackets db41a36af4 `Genome::process` first implementation with a single `output` + must add 2 layers to have the last one as the `output` 2016-10-27 00:19:28 +02:00
xdrm-brackets 20382821ca Done Genome serialization/unserialization [Genome::serialize()] + [Genome::unserialize(String)] 2016-10-26 18:07:52 +02:00
xdrm-brackets 7cacdb3b63 Corrections + Genome::mutation(threshold) 2016-10-26 17:53:56 +02:00
xdrm-brackets 3c57a83e41 Correcting genome inheritance & crossover (wrong number of neurons > now its neurons*layers) 2016-10-26 17:24:51 +02:00
xdrm-brackets fd9bcefed7 Created `Genome`'s `constructor` + finished `NeuralNetwork`'s `storage` and `attributes` management (+ serialization) + now it remains the hard & cool part -> genetic algorithm 2016-10-26 17:15:00 +02:00
xdrm-brackets 228d1d7913 Project structure + class bases and use cases for the Neural Network + FileManager to manage large files 2016-10-26 00:52:28 +02:00