Matmul rework

[ADD] support for Multi-dimensional Matrix Multiplication, following numpy MatMul operation specificities.

In the context of Multi-dimensional MatMul, we consider an N-Dim Tensor as a stack of 2-D matrices having as shape the last two dimensions of the Tensor.

From Numpy documentation:

The behavior depends on the arguments in the following way.

• If both arguments are 2-D they are multiplied like conventional matrices.
• If either argument is N-D, N > 2, it is treated as a stack of matrices residing in the last two indexes and broadcast accordingly.
• If the first argument is 1-D, it is promoted to a matrix by prepending a 1 to its dimensions. After matrix multiplication the prepended 1 is removed.
• If the second argument is 1-D, it is promoted to a matrix by appending a 1 to its dimensions. After matrix multiplication the appended 1 is removed.

src/operator/MatMul.cpp

@@ -32,6 +32,11 @@ void Aidge::MatMul_Op::computeOutputDims() {
         std::vector<std::size_t> dims0 = getInput(0)->dims();
         std::vector<std::size_t> dims1 = getInput(1)->dims();
 
+        // keep second-to-last dimension of dims0
+        const bool keepDim0 = dims0.size() > 1;
+        // keep last dimension of dims1
+        const bool keepDim1 = dims1.size() > 1;
+
         if (dims0.size() == 1) {
             dims0.insert(dims0.cbegin(), 1);
         }
@@ -42,10 +47,10 @@ void Aidge::MatMul_Op::computeOutputDims() {
 
 
         if (dims0.size() > dims1.size()) {
-                dims1.insert(dims1.cbegin(), dims0.begin(), dims0.end() - dims1.size());
+            dims1.insert(dims1.cbegin(), dims0.begin(), dims0.end() - dims1.size());
         }
         else if (dims1.size() > dims0.size()) {
-                dims0.insert(dims0.cbegin(), dims1.begin(), dims1.end() - dims0.size());
+            dims0.insert(dims0.cbegin(), dims1.begin(), dims1.end() - dims0.size());
         }
 
         AIDGE_ASSERT(dims0[dims_size-1] == dims1[dims_size-2], "Incompatible matrices sizes.");
@@ -56,11 +61,10 @@ void Aidge::MatMul_Op::computeOutputDims() {
             outDims[i] = std::max(dims0[i], dims1[i]);
         }
 
-        // keep second-to-last dimension of dims0
-        if (dims0.size() > 1)
+        // use keepDim0 instead of dims0.size() because dims0 has been modified
+        if (keepDim0)
             outDims.push_back(dims0[dims_size-2]);
-        // keep last dimension of dims1
-        if (dims1.size() > 1)
+        if (keepDim1)
             outDims.push_back(dims1[dims_size-1]);
 
         mOutputs[0]->resize(outDims);