Search Results for "sparse matrices"

The blocksparse repository provides efficient GPU kernels (TensorFlow custom ops) for block-sparse matrix multiplication and convolution operations. The idea is to exploit block-level sparsity — i.e. treat matrices or weight tensors as composed of blocks, many of which may be zero or unused — to save compute and memory when sparsity patterns are structured. This is particularly useful in models like Sparse Transformers, where attention matrices or intermediate layers may adopt block-sparse patterns to scale better. ...

The package yaUFget is yet another interface to the collection of sparse matrices at the University of Florida for Python. In contrast with the official Python interface it can be easily used as a package and can be imported to any place, obtains the updated index of matrices from the collection's web site and does not require MATLAB. The functionality of the MATLAB's interface UFget was preserved.