Torch is not in active developement. The functionality provided by the C backend of Torch, which are the TH, THNN, THC, THCUNN libraries is actively extended and re-written in the ATen C++11 library (source, mirror). ATen exposes all operators you would expect from torch7, nn, cutorch, and cunn directly in C++11 and includes additional support for sparse tensors and distributed operations. It is to note however that the API and semantics of the backend libraries in Torch-7 are different from the semantice provided by ATen. For example ATen provides numpy-style broadcasting while TH* dont. For information on building the forked Torch-7 libraries in C, refer to "The C interface" in pytorch/aten/src/README.md.
Torch7 community support can be found at the following locations. However Torch7 has a much smaller active community than it used to. If you have questions about the C backend of Torch-7, you can try asking in the PyTorch communication channels, as the developers are familiar with it.
- Questions, Support, Install issues: Google groups
- Reporting bugs: torch7 nn cutorch cunn optim threads
- Hanging out with other developers and users (strictly no install issues, no large blobs of text): Gitter Chat
Torch Package Reference Manual
Torch is the main package in Torch7 where data structures for multi-dimensional tensors and mathematical operations over these are defined. Additionally, it provides many utilities for accessing files, serializing objects of arbitrary types and other useful utilities.
- Tensor Library
- File I/O Interface Library
- Useful Utilities
- Timer provides functionality for measuring time.
- Tester is a generic tester framework.
- CmdLine is a command line argument parsing utility.
- Random defines a random number generator package with various distributions.
- Finally useful utility functions are provided for easy handling of torch tensor types and class inheritance.