search

MIT-REALM / neural_clbf

Toolkit for learning controllers based on robust control Lyapunov barrier functions
BSD 3-Clause "New" or "Revised" License
124 stars 43 forks source link

Make Gurobi an optional flag. #8

Closed rallen10 closed 1 year ago

rallen10 commented 1 year ago

I am trying to get up and running with this codebase and I immediately run into a licensing problem when I try to run:

python neural_clbf/training/train_inverted_pendulum.py
.
.
.
  File "src/gurobipy/env.pxi", line 67, in gurobipy.Env.__init__
  File "src/gurobipy/gurobi.pxi", line 32, in gurobipy.gurobi._getdefaultenv
  File "src/gurobipy/env.pxi", line 67, in gurobipy.Env.__init__
gurobipy.GurobiError: License expired 2022-01-13

Since CVX is already a dependency, it seems that you could just set an optional flag in clf_controller.py to use gurobi if desired

Gurobi's academic licensing has several hoops to jump through in order to obtain. Not only does it require an academic affiliation (e.g. email), you actually have to be connected to a university-affiliated IP for the license install to work.

rallen10 commented 1 year ago

Proposed fix here: https://github.com/MIT-REALM/neural_clbf/pull/9

dawsonc commented 1 year ago

Thanks for starting a PR on this; we can move discussion over to #9

dawsonc commented 1 year ago

Closed by #9