Scalar of Type Float for Code Generation

[plancherb1]

Hi, I've been trying to use the code generation features and have found that for some reason I can only get them to support a Scalar of type double. Is there any way to get them to accept a Scalar of type float?

For example setting the type parameter T in the below code snippet to double works fine, but setting it to float results in a compilation error stating that there is: no matching constructor for initialization of 'CodeGenRNEA<float> which is I think caused by this other error: note: candidate constructor (the implicit copy constructor) not viable: no known conversion from 'pinocchio::Model' (aka 'ModelTpl<double>') to 'const pinocchio::CodeGenRNEA<float>' for 1st argument struct CodeGenRNEA : public CodeGenBase<_Scalar>.

Thanks in advance, I assume I am missing something simple!

// Import URDF model and set gravity
Model model;
pinocchio::urdf::buildModel(urdf_filepath,model);
model.gravity.linear(Eigen::Vector3d(0,0,-9.81));

// generate the code_gen
CodeGenRNEA<T> rnea_code_gen(model);
rnea_code_gen.initLib();
rnea_code_gen.loadLib();

[jcarpent]

Hi @plancherb1! You just need to cast your input model as follows and similarly to Eigen:

CodeGenRNEA<T> rnea_code_gen(model.cast<T>());

[jcarpent]

I've closed the issue as the issue seems to be solved. Feel free to reopen it if needed