meta-torch-core-free 20171127 source package in Ubuntu


meta-torch-core-free (20171127) unstable; urgency=medium

  * Allow outputs to stderr to avoid CI failure.
  * Recommends lua-torch-cutorch.
  * Drop the meaningless versioning scheme. Use the last-update
    timestamp among all dependency package repositories of upstream
    as the metapackage version.

 -- Mo Zhou <email address hidden>  Sun, 24 Jun 2018 11:30:02 +0000

Upload details

Uploaded by:
Debian Science Team on 2018-06-24
Uploaded to:
Original maintainer:
Debian Science Team
Medium Urgency

See full publishing history Publishing

Series Pocket Published Component Section
Eoan release on 2019-04-18 universe misc
Disco release on 2018-10-30 universe misc
Cosmic release on 2018-06-24 universe misc


File Size SHA-256 Checksum
meta-torch-core-free_20171127.dsc 2.0 KiB d689392f633c596447e68a49a9a9d7e8a2b15b2e5ada45c975f26d140de351b3
meta-torch-core-free_20171127.tar.xz 3.8 KiB 081102efc04389276b5053aeddfe1a24cd196ef8a76a0a1edd98a0bf8038c694

Available diffs

No changes file available.

Binary packages built by this source

torch-core-free: Scientific Computing Framework For Luajit (Core Components)

 Torch is a scientific computing framework with wide support for machine
 learning algorithms that puts GPUs first. It is easy to use and efficient,
 thanks to an easy and fast scripting language, LuaJIT, and an underlying
 C/CUDA implementation.
 A summary of core features:
  * a powerful N-dimensional array
  * lots of routines for indexing, slicing, transposing, ...
  * amazing interface to C, via LuaJIT
  * linear algebra routines
  * neural network, and energy-based models
  * numeric optimization routines
  * Fast and efficient GPU support
  * Embeddable, with ports to iOS, Android and FPGA backends
 The goal of Torch is to have maximum flexibility and speed in building
 your scientific algorithms while making the process extremely simple.
 Torch comes with a large ecosystem of community-driven packages in
 machine learning, computer vision, signal processing, parallel
 processing, image, video, audio and networking among others, and
 builds on top of the Lua community.
 At the heart of Torch are the popular neural network and optimization
 libraries which are simple to use, while having maximum flexibility
 in implementing complex neural network topologies. You can build
 arbitrary graphs of neural networks, and parallelize them over CPUs
 and GPUs in an efficient manner.
 This package is a metapackage, which pulls the following core and *free*
 modules for you: cwrap, paths, sys, xlua, torch7, nn, graph, nngraph,
 optim, sundown, dok, trepl, image.
 Note that cutorch (CUDA backend for torch) and cunn (CUDA backend for
 neural network) are not present in this metapacakge - they will be
 shipped in the torch-core-contrib metapackage in the future.