Neural network on sparse matrix

I have a large sparse matrix. I want to perform neural network on this sparse matrix. Is there support for doing so? If so, how?

Hi @Mckenzie

Create normal sparse tensor for your sparse matix like this commad

# Suppose your sparse matrix is stored in COO format (coordinates, values, shape)
sparse_matrix = tf.sparse.SparseTensor(indices=coordinates, values=values, dense_shape=shape)


# Convert COO format to PyTorch sparse tensor
sparse_matrix = torch.sparse_coo_tensor(coordinates, values, size=shape)

And After that apply your NN normally without changes

Any idea how to do the same in R?