Hi Satra, for a training set of shape (n_samples, n_features) the required memory should be O(n_samples * max(n_samples, n_features, n_components)).
The following additional memory is allocated: * matrix of pairwise distances 'distances' (n_samples, n_samples) * conditional probabilities 'conditional_P' (n_samples, n_samples) * joint probabilities 'P' (n_samples, n_samples) * matrix of pairwise distances in the embedded space 'n' (n_samples * (n_samples - 1)) * Student's t-distribution 'Q' (n_samples * (n_samples - 1)) * Kullback-Leibler divergence 'kl_divergence' (n_samples * (n_samples - 1)) * 'PQd' (n_samples, n_samples) * 'grad' (n_samples * n_components) * embedded data 'self.embedding_' (n_samples, n_components) Best regards, Alexander > hi folks, > > does anyone know what the max memory usage is of TSNE in scikit learn? > for a n x n dense matrix, how much memory is required? > > cheers, > > satra ------------------------------------------------------------------------------ Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server from Actuate! Instantly Supercharge Your Business Reports and Dashboards with Interactivity, Sharing, Native Excel Exports, App Integration & more Get technology previously reserved for billion-dollar corporations, FREE http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk _______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
