def on_monitor(self, model, dataset, algorithm): model = algorithm.model epoch = algorithm.monitor.get_epochs_seen() fn = self.base + str(epoch) + '.' + self.format outfn = os.path.join(self.dir, fn) pv = get_weights_report.get_weights_report(model=model) pv.save(outfn)
def on_monitor(self, model, dataset, algorithm): """ Looks whether the model performs better than earlier. If it's the case, saves the model. Parameters ---------- model : pylearn2.models.model.Model model.monitor must contain a channel with name given by self.channel_name dataset : pylearn2.datasets.dataset.Dataset not used algorithm : TrainingAlgorithm not used """ monitor = model.monitor channels = monitor.channels channel = channels[self.channel_name] val_record = channel.val_record new_cost = self.coeff * val_record[-1] if new_cost < self.best_cost: self.best_cost = new_cost serial.save(self.save_path, model, on_overwrite = 'backup') # XXX: [Kien] Save best filters. pv = get_weights_report.get_weights_report(model = model, dataset = dataset) pv.save('best_filters.png')
def main(): parser = OptionParser() parser.add_option("--rescale", dest='rescale', type='string', default="individual") parser.add_option("--out", dest="out", type='string', default=None) parser.add_option("--border", dest="border", action="store_true", default=False) options, positional = parser.parse_args() assert len(positional) == 1 path, = positional rescale = options.rescale border = options.border pv = get_weights_report.get_weights_report(model_path=path, rescale=rescale, border=border) if options.out is None: pv.show() else: pv.save(options.out)
def my_show_weights(model_path, rescale='individual', border=False, out=None): pv = get_weights_report.get_weights_report(model_path=model_path, rescale=rescale, border=border) if out is None: pv.show() else: pv.save(out)
def showWeights(model_name, model, rescale="individual", border=None, out=None): pv = get_weights_report.get_weights_report(model_path=model_name, rescale=rescale, border=border) if out is None: pv.show() else: pv.save(out)
def showWeights(model_name, model, rescale="individual", border=None, out=None): pv = get_weights_report.get_weights_report(model_path=model_name, rescale=rescale, border=border) if out is None: pv.show() else: pv.save(out)
def main(): parser = argparse.ArgumentParser() parser.add_argument("--rescale", default="individual") parser.add_argument("--out", default=None) parser.add_argument("--border", action="store_true", default=False) parser.add_argument("path") options = parser.parse_args() pv = get_weights_report.get_weights_report(model_path = options.path, rescale = options.rescale, border = options.border) if options.out is None: pv.show() else: pv.save(options.out)
def main(): parser = argparse.ArgumentParser() parser.add_argument("--rescale", default="individual") parser.add_argument("--out", default=None) parser.add_argument("--border", action="store_true", default=False) parser.add_argument("path") options = parser.parse_args() pv = get_weights_report.get_weights_report(model_path=options.path, rescale=options.rescale, border=options.border) if options.out is None: pv.show() else: pv.save(options.out)
def train(self, dataset): # Call LBFGS self.XS = theano.shared(dataset.X) self.costfn = self.cost(self.model, self.XS) vec = np.ones(29144) #from scipy.optimize import fmin_bfgs from scipy.optimize import fmin_l_bfgs_b vecstar = fmin_l_bfgs_b(f,x0=self.model_to_vector(), fprime=fprime, args=(self,), factr=1e5) opt = vecstar[0] self.update_model(opt) from pylearn2.gui import get_weights_report pv = get_weights_report.get_weights_report(model=self.model) pv.save('output.png') fprime(vec, self) 1/0 pass
def main(): parser = OptionParser() parser.add_option("--rescale",dest='rescale',type='string',default="individual") parser.add_option("--out",dest="out",type='string',default=None) parser.add_option("--border", dest="border", action="store_true",default=False) options, positional = parser.parse_args() assert len(positional) == 1 path ,= positional rescale = options.rescale border = options.border pv = get_weights_report.get_weights_report(model_path = path, rescale = rescale, border = border) if options.out is None: pv.show() else: pv.save(options.out)
def show_weights(model_path, rescale="individual", border=False, out=None): """ Show or save weights to an image for a pickled model Parameters ---------- model_path : str Path of the model to show weights for rescale : str WRITEME border : bool, optional WRITEME out : str, optional Output file to save weights to """ pv = get_weights_report.get_weights_report(model_path=model_path, rescale=rescale, border=border) if out is None: pv.show() else: pv.save(out)
def show_weights(model_path, rescale="individual", border=False, out=None): """ Show or save weights to an image for a pickled model Parameters ---------- model_path : str Path of the model to show weights for rescale : str WRITEME border : bool, optional WRITEME out : str, optional Output file to save weights to """ pv = get_weights_report.get_weights_report(model_path=model_path, rescale=rescale, border=border) if out is None: pv.show() else: pv.save(out)
#!/usr/bin/env python __authors__ = "Ian Goodfellow" __copyright__ = "Copyright 2010-2012, Universite de Montreal" __credits__ = ["Ian Goodfellow"] __license__ = "3-clause BSD" __maintainer__ = "Ian Goodfellow" __email__ = "goodfeli@iro" import sys from pylearn2.gui import get_weights_report import warnings warnings.warn("make_weights_image.py is deprecated. Use show_weights.py with" " the --out flag. make_weights_image.py may be removed on or after " "2014-08-28.") if __name__ == "__main__": print 'loading model' path = sys.argv[1] print 'loading done' rescale = True if len(sys.argv) > 2: rescale = eval(sys.argv[2]) pv = get_weights_report.get_weights_report(path, rescale) pv.save(sys.argv[1]+'.png')
WRITEME """ __authors__ = "Ian Goodfellow" __copyright__ = "Copyright 2010-2012, Universite de Montreal" __credits__ = ["Ian Goodfellow"] __license__ = "3-clause BSD" __maintainer__ = "Ian Goodfellow" __email__ = "goodfeli@iro" import sys from pylearn2.gui import get_weights_report import warnings warnings.warn( "make_weights_image.py is deprecated. Use show_weights.py with" " the --out flag. make_weights_image.py may be removed on or after " "2014-08-28.") if __name__ == "__main__": print 'loading model' path = sys.argv[1] print 'loading done' rescale = True if len(sys.argv) > 2: rescale = eval(sys.argv[2]) pv = get_weights_report.get_weights_report(path, rescale) pv.save(sys.argv[1] + '.png')
#!/usr/bin/env python #usage: show_weights.py model.pkl from pylearn2.gui import get_weights_report from optparse import OptionParser parser = OptionParser() parser.add_option("--rescale",dest='rescale',type='string',default="individual") parser.add_option("--out",dest="out",type='string',default=None) parser.add_option("--border", dest="border", action="store_true",default=False) options, positional = parser.parse_args() assert len(positional) == 1 path ,= positional rescale = options.rescale border = options.border pv = get_weights_report.get_weights_report(model_path = path, rescale = rescale, border = border) if options.out is None: pv.show() else: pv.save(options.out)
from optparse import OptionParser parser = OptionParser() parser.add_option("--rescale", dest='rescale', type='string', default="individual") parser.add_option("--out", dest="out", type='string', default=None) parser.add_option("--border", dest="border", action="store_true", default=False) options, positional = parser.parse_args() assert len(positional) == 1 path, = positional rescale = options.rescale border = options.border pv = get_weights_report.get_weights_report(model_path=path, rescale=rescale, border=border) if options.out is None: pv.show() else: pv.save(options.out)
from pylearn2.utils import serial import sys model_path, kmeans_path = sys.argv[1:] print 'loading model' model = serial.load(model_path) print 'loading kmeans' kmeans = serial.load(kmeans_path) from galatea.s3c.s3c_dataset import S3C_Dataset from pylearn2.config import yaml_parse print 'loading dataset' raw = yaml_parse.load(model.dataset_yaml_src) print 'making transformer dataset' dataset = S3C_Dataset(raw = raw, transformer = model) from pylearn2.gui.get_weights_report import get_weights_report print 'making weights report' pv = get_weights_report(model = kmeans, dataset = dataset) pv.show()