MonteA machine learning library written in Python | |
Download |
Monte Ranking & Summary
Advertisement
- License:
- Freeware
- Price:
- FREE
- Publisher Name:
- Roland Memisevic
- Publisher web site:
- Operating Systems:
- Mac OS X
- File Size:
- 153 KB
Monte Tags
Monte Description
A machine learning library written in Python Monte (python) is a free Python framework for building gradient based learning machines, like neural networks, logistic regression, conditional random fields, etc. Monte contains modules (that hold parameters, a cost-function and a gradient-function) and trainers (that can adapt a module's parameters by minimizing its cost-function on training data).Modules are usually composed of other modules, which can in turn contain other modules, etc. Gradients of decomposable systems like these can be computed with back-propagation. Requirements: · Python 2.4 or later · NumPy 1.0 or later · SciPy 0.5.1 or later · matplotlib 0.87 or later What's New in This Release: · Major overhaul, bug fixes. Change in directory structure: New directory models/ contains trainable models, contrastive is now a sub-package of models/ some important changes: · util.onehot and util.unhot deal with arrays of any shape, one-hat encoding always along the LAST dimension. · util.logsumexp deals with arrays of any shape, works by default along the LAST dimension. · new module gym.paralleltrainer added. currently contains one class that constitutes a parallel trainer using threading. · bug fix in models.contrastive.rbm that fixes a problem with cditerations>1. · several new models added, that use bp.neuralnet-components.
Monte Related Software