sparsity

Modules

modelopt.torch.kernels.sparsity.attention	Kernel integrations for sparse attention: Triton FA and diffusers/LTX backends.
modelopt.torch.kernels.sparsity.gemm	Sparsity GEMM kernels (placeholder for future implementations).

Sparsity kernels: attention (Triton skip-softmax backends) and gemm (placeholder).