Handle (create) model and jacobian in my custom factor

[mohammadreza dindarloo]

hello everyone. 
untill now I tried to develop some custom factors. for all of them I parametricaly calculate my model base on my variables(i. e. h(q) that q is my variables vector) and by partial differential jacobians were extracted(i.e. H). So by this method I created my error vector in errorEvaluating method (E(q) = h(q) - m (m is my measurement)). Now in another project I want to calculate my model base on variabled (E(q)) and then by diff from that take jacobian (H) and create my custom factor. Now my problem it that my model (h(q)) is very big and I can't calculate it easily and then finding H. Base on my research the way of creating custom factor to make my own optimization it this but unfortunaly even maple can't handle equation because of big matrix. Now I want to know is my road map of creating custom factor true? should I find all the model base of variables in one equation?

 if it is possible please give me some refrences or hints to search about solution.
thanks all,
best regards,
M.R. Dindarloo

[mohammadreza dindarloo]

hi everyone
I found a good solution in my question. I used symforce python library for computing my jacobian and model. it you had a same problem take a look symforce.

best wishes
Mreza Dindarloo

[Dellaert, Frank]

Symforce is very cool! Glad you were able to find a solution.

Best

Frank Dellaert

http://www.cc.gatech.edu/~dellaert

On Aug 26, 2023, at 8:19 AM, mohammadreza dindarloo <mohammadre...@gmail.com> wrote:

hi everyone

--
You received this message because you are subscribed to the Google Groups "gtsam users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gtsam-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gtsam-users/04039f41-e84e-498e-ae73-e2eb76f24922n%40googlegroups.com.