Skip to main content
This is a pre-production deployment of Warehouse. Changes made here affect the production instance of PyPI (
Help us improve Python packaging - Donate today!

Fast, flexible and fun neural networks.

Project Description


The full documentation is at


0.5 (2015-12-01)

Changed Behaviour

  • examples now run on CPU by default
  • added and to help with data preparation
  • SigmoidCE and SquaredDifference layers now outputs a loss for each dimension instead of summing over features.
  • SquaredDifference layer does no longer scale by one half.
  • Added a SquaredLoss layer that computes half the squared difference and has an interface that is compatible with the SigmoidCE and SigmoidCE layers.
  • Output probabilities renamed to predictions in SigmoidCE and SigmoidCE layers.

New Features

  • added a use_conv option to
  • added criterion option to brainstorm.hooks.EarlyStopper hook
  • added function that returns information about the network as a string
  • added function that applies a network to some data and saves a set of requested buffers.
  • brainstorm.layers.mask layer now supports masking individual features
  • added brainstorm.hooks.StopAfterThresholdReached hook


  • EarlyStopper now works for any timescale and interval
  • Recurrent, Lstm, Clockwork, and ClockworkLstm layers now accept inputs of arbitrary shape by implicitly flattening them.
  • several fixes to make building the docs easier
  • some performance improvements of NumpyHandler operations binarize_t and index_m_by_v
  • sped up tests
  • several improvements to installation scripts


  • fixed sqrt operation for PyCudaHandler. This should fix problems with BatchNormalization on GPU.
  • fixed a bug for task_type=’regression’ in and
  • removed defunct name argument from input layer
  • fixed a crash when applying brainstorm.hooks.SaveBestNetwork to rolling_training loss
  • various minor fixes of the brainstorm.hooks.BokehVisualizer
  • fixed a problem with sum_t operation in brainstorm.handlers.PyCudaHandler
  • fixed a blocksize problem in convolutional and pooling operations in brainstorm.handlers.PyCudaHandler

0.5b0 (2015-10-25)

  • First release on PyPI.
Release History

Release History

This version
History Node


History Node


History Node


Download Files

Download Files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
brainstorm-0.5.tar.gz (246.3 kB) Copy SHA256 Checksum SHA256 Source Dec 1, 2015

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting