First release.

• core contributors, in alphabetical order:
  • Eric Battenberg (@ebattenberg)
  • Sander Dieleman (@benanne)
  • Daniel Nouri (@dnouri)
  • Eben Olson (@ebenolson)
  • Aäron van den Oord (@avdnoord)
  • Colin Raffel (@craffel)
  • Jan Schlüter (@f0k)
  • Søren Kaae Sønderby (@skaae)
• extra contributors, in chronological order:
  • Daniel Maturana (@dimatura): documentation, cuDNN layers, LRN
  • Jonas Degrave (@317070): get_all_param_values() fix
  • Jack Kelly (@JackKelly): help with recurrent layers
  • Gábor Takács (@takacsg84): support broadcastable parameters in lasagne.updates
  • Diogo Moitinho de Almeida (@diogo149): MNIST example fixes
  • Brian McFee (@bmcfee): MaxPool2DLayer fix
  • Martin Thoma (@MartinThoma): documentation
  • Jeffrey De Fauw (@JeffreyDF): documentation, ADAM fix
  • Michael Heilman (@mheilman): NonlinearityLayer, lasagne.random
  • Gregory Sanders (@instagibbs): documentation fix
  • Jon Crall (@Erotemic): check for non-positive input shapes
  • Hendrik Weideman (@hjweide): set_all_param_values() test, MaxPool2DCCLayer fix
  • Kashif Rasul (@kashif): ADAM simplification
  • Peter de Rivaz (@peterderivaz): documentation fix