[torch-mlir][sparse] register sparse tensor dialect for all rewriting (llvm#3918)

We incorrectly relied on the fact that StableHLO registers the sparse
tensor dialect, but when building for e.g. just LinAlg, the dependency
was missing. This fixes this shortcoming.

FIXES: llvm#3816

Author: aartbik

lib/InitAll.cpp

@@ -16,6 +16,7 @@
 #include "mlir/Dialect/MLProgram/IR/MLProgram.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/SCF/IR/SCF.h"
+#include "mlir/Dialect/SparseTensor/IR/SparseTensor.h"
 #include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/Dialect/Tensor/IR/TensorInferTypeOpInterfaceImpl.h"
 #include "mlir/Dialect/Tosa/IR/TosaOps.h"
@@ -52,7 +53,8 @@ void mlir::torch::registerOptionalInputDialects(
     mlir::DialectRegistry &registry) {
   registry.insert<complex::ComplexDialect, linalg::LinalgDialect,
                   memref::MemRefDialect, ml_program::MLProgramDialect,
-                  scf::SCFDialect, tensor::TensorDialect, tosa::TosaDialect>();
+                  scf::SCFDialect, sparse_tensor::SparseTensorDialect,
+                  tensor::TensorDialect, tosa::TosaDialect>();
 }
 
 void mlir::torch::registerAllPasses() {