How to use it ? #5

Activity

1. arpit1997 commented on Oct 30, 2016

   

   arpit1997

   on Oct 30, 2016

   More actions

   @IITtense Well it seems pretty easy to use for a data set and training it is easy and fast because it does not require you to tune the parameters. check this out for details and here is some examples for this.

   >>> from hpelm import ELM
   >>> elm = ELM(X.shape[1], T.shape[1])
   >>> elm.add_neurons(20, "sigm")
   >>> elm.add_neurons(10, "rbf_l2")
   >>> elm.train(X, T, "LOO")
   >>> Y = elm.predict(X)
   

   • hpelm.ELM for in-memory computations (dataset fits into RAM)
   • hpelm.HPELM for out-of-memory computations (dataset on disk in HDF5 format)

   I hope this gives some idea. 😸 😸

   👍React with 👍2Reacted by Renato Sousa and namedysx
2. alyato commented on Jan 18, 2018

   

   alyato

   on Jan 18, 2018

   More actions

   @arpit1997
   Are there some explame to test?
   Im using the CNN to train my dataset now.But i want to compare CNN with the ELM.