What is ttnn?

ttnn is a library that provides a user-friendly interface to operations that run on TensTorrent’s hardware using tt-metal programming model.

ttnn is designed to be intuitive to an user that is familiar with PyTorch.

ttnn’s primary dependency is tt_lib which provides the implementation for all of the operations used by ttnn.

We trust that this library will be valuable to helping you on your journey to take full advantage of our devices!

Key features of ttnn

Key features of ttnn:

• Support for N-D tensors.

• Intuitive way of converting between ttnn.ROW_MAJOR_LAYOUT and ttnn.TILE_LAYOUT using ttnn.to_layout

• Stable APIs.

• The computation graph of ttnn operations can be traced and then visualized or used in any other way. The graph is networkx compatible. Refer to ttnn Tracer for examples

• Infrastructure for converting parameters and some sub-modules from a torch.nn.Module object. This infrastructure supports caching of the converted parameters which could significantly speed up repeated runs.

• Ability to compare the result of each operation to the equivalent PyTorch operation. Very useful for debugging.