jimpil / enclog Goto Github PK
View Code? Open in Web Editor NEWClojure wrapper for Encog (v3) (Machine-Learning framework that specialises in neural-nets)
enclog's People
Contributors
Stargazers
Watchers
Forkers
clojens teachshits hellcoderz smartrevolution spradnyesh weaver-viii chunsj stjordanis standardgalacticenclog's Issues
Support for Clojure 1.8
Is this supported in Clojure 1.8?
Encog 3.2 breaks enclog, it changes a lot of structure.
At least NeuralGeneticAlgorithm and NEATTraining
For NeuralGeneticAlgoritm, I found there is now org.encog.ml.ea.train.EvolutionaryAlgorithm but I don't know if it can be used analogously.
there is a bug in v0.6.5
(ns resilient.core
;;(:use incanter.core)
(:use [enclog nnets training])
(:gen-class))
(defn round-output
"Round outputs to nearest integer."
[output]
(mapv #(Math/round ^Double %) output))
(def dataset
(let [xor-input [[0.0 0.0] [1.0 0.0] [0.0 1.0] [1.0 1.0]]
xor-ideal [[0.0] [1.0] [1.0] [0.0]]]
(data :basic-dataset xor-input xor-ideal)))
(defn train-network [network data trainer-algo]
(let [trainer (trainer trainer-algo
:network network
:training-set data)]
(train trainer 0.01 10000 [])))
(def elman-network (network (neural-pattern :elman)
:activation :sigmoid
:input 2
:output 1
:hidden [3]))
(defn run-network [network input]
(let [input-data (data :basic input)
output (.compute network input-data)
output-vec (.getData output)]
(round-output output-vec)))
(def EN (train-network elman-network dataset :resilient-prop))as i configure dependencies in project.clj
:dependencies [[org.clojure/clojure "1.6.0"]
[cc.artifice/clj-ml "0.4.0"]
[org.encog/encog-core "3.2.0"]
[enclog "0.6.5"]
[org.clojure/tools.namespace "0.2.5"]
[incanter "1.5.5"]
[clatrix "0.3.0"]
[net.mikera/core.matrix "0.27.0"]]and the log generate from train-network function is that
Iteration # 9,978 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,979 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,980 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,981 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,982 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,983 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,984 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,985 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,986 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,987 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,988 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,989 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,990 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,991 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,992 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,993 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,994 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,995 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,996 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,997 Error: 10.879860% Target-Error: 1.000000%
Iteration # 9,998 Error: 20.051045% Target-Error: 1.000000%
Iteration # 9,999 Error: 10.879860% Target-Error: 1.000000%
Iteration # 10,000 Error: 20.051045% Target-Error: 1.000000%
as you can see after iterate to 10000 times it still get a large error...
but when i change encog-core to 3.1.0 and enclog to 0.6.3 the log when i train the elman network shows that
Iteration # 146 Error: 6.437994% Target-Error: 1.000000%
Iteration # 147 Error: 6.162065% Target-Error: 1.000000%
Iteration # 148 Error: 5.905370% Target-Error: 1.000000%
Iteration # 149 Error: 5.685170% Target-Error: 1.000000%
Iteration # 150 Error: 8.820329% Target-Error: 1.000000%
Iteration # 151 Error: 6.344310% Target-Error: 1.000000%
Iteration # 152 Error: 6.236660% Target-Error: 1.000000%
Iteration # 153 Error: 6.115702% Target-Error: 1.000000%
Iteration # 154 Error: 5.980763% Target-Error: 1.000000%
Iteration # 155 Error: 5.831186% Target-Error: 1.000000%
Iteration # 156 Error: 5.666209% Target-Error: 1.000000%
Iteration # 157 Error: 5.484842% Target-Error: 1.000000%
Iteration # 158 Error: 5.285799% Target-Error: 1.000000%
Iteration # 159 Error: 5.067628% Target-Error: 1.000000%
Iteration # 160 Error: 4.829173% Target-Error: 1.000000%
Iteration # 161 Error: 4.570456% Target-Error: 1.000000%
Iteration # 162 Error: 4.294153% Target-Error: 1.000000%
Iteration # 163 Error: 4.008254% Target-Error: 1.000000%
Iteration # 164 Error: 3.732056% Target-Error: 1.000000%
Iteration # 165 Error: 3.511699% Target-Error: 1.000000%
Iteration # 166 Error: 3.360398% Target-Error: 1.000000%
Iteration # 167 Error: 3.504944% Target-Error: 1.000000%
Iteration # 168 Error: 4.968425% Target-Error: 1.000000%
Iteration # 169 Error: 0.958797% Target-Error: 1.000000%
when iterate to 169 times and the algorithm converged
so anyone can show me how to fix the issue in enclog 0.6.5 ?
what's more i think resilient propgate algorithm did not perform as well as backpropgate algorithm :(
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.