Le 21 mars 2012 01:21, David Marek <[email protected]> a écrit : > Hi > > I think I was a little confused, I'll try to summarize what I > understand is needed: > > * the goal is to have multilayer perceptron with stochastic gradient > descent and maybe other learning algorithms > * basic implementation of sgd already exists > * existing implementation is missing some loss-functions, we need to > use some from sgd-fast and implement others > * existing implementation is slow, the final implementation should be > written in Cython > > Am I correct?
It seems so. > I have experimented with Cython and realized that there will be work > needed to make the mlp faster. Preallocating the arrays to store that activations and using minibatches so as to perform SIMD optimized and thread parallelized DGEMM blas calls rather than single sample DGEMV is probably also very important for perfomance. There is a trade-off between hardward speedup (the larger the minibatches the better) and algorithmic speedup (the smaller the minibatches the better for real online SGD). Also please be sure to read: http://yann.lecun.com/exdb/publis/#lecun-98b -- Olivier http://twitter.com/ogrisel - http://github.com/ogrisel ------------------------------------------------------------------------------ This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure _______________________________________________ Scikit-learn-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
