Giovanni De Toni

> Just realised, the solution is an observer Just to understand better, do you need to expose just computed values or directly a function which needs to be called by...

> Hmmm, I guess this is why the random forest meta example sometimes fails locally... Have you tried to run it in gdb yet? Nope, I still need to check...

The stacktrace is the following: ``` #0 0x00007ffff58156b4 in std::__shared_count::__shared_count (this=0x7fffce914c38, __r=...) at /usr/include/c++/7/bits/shared_ptr_base.h:849 #1 0x00007ffff57cb20d in std::__shared_ptr::__shared_ptr ( this=0x7fffce914c30, __r=...) at /usr/include/c++/7/bits/shared_ptr_base.h:1147 #2 0x00007ffff576818b in std::shared_ptr::shared_ptr (this=0x7fffce914c30, __r=...) at...

Thank you @gf712 for the insight!

If we add `m_root` as a parameter we will get the following messages once we try to deserialize: ``` [06/09/20 10:45:49 error] Could not create 'BinaryTreeMachineNode' class [06/09/20 10:45:49 warning]...

As I mentioned in the email, also this lib here could be used as an inspiration https://github.com/glm-tools/pyglmnet

> > linearmachine? > > if you mean inheriting from LinearMachine well the issue comes that it already has things like `bias` that aren't applicable at all -to my knowledge-...

The bias is usually represented by `beta_0`. The epsilon means a noise value.

@gf712 yes, the example you wrote is exactly what I had in mind :) Then if the user wants to observe the parameter `oob_error`, once the observer captures the lambda,...

> > The use case is that a user has applied a filter to the observer to choose what to store....in that case we dont always want to compute the...