Do not collect training data during benchmarks

miplearn/benchmark.py

@@ -27,7 +27,8 @@ class BenchmarkRunner:
         for (name, solver) in self.solvers.items():
             results = solver.parallel_solve(instances,
                                             n_jobs=n_jobs,
-                                            label=name)
+                                            label=name,
+                                            collect_training_data=False)
             for i in range(len(instances)):
                 wallclock_time = None
                 for key in ["Time", "Wall time", "Wallclock time"]:

miplearn/solvers.py

@@ -55,9 +55,6 @@ class LearningSolver:
         self.internal_solver = self.internal_solver_factory()
         self.is_persistent = hasattr(self.internal_solver, "set_instance")
         
-    def _clear(self):
-        self.internal_solver = None
-
     def solve(self, instance, tee=False):
         model = instance.to_model()
         
@@ -80,13 +77,20 @@ class LearningSolver:
         
         return solve_results
                 
-    def parallel_solve(self, instances, n_jobs=4, label="Solve"):
-        self._clear()
+    def parallel_solve(self,
+                       instances,
+                       n_jobs=4,
+                       label="Solve",
+                       collect_training_data=True,
+                      ):
+        self.internal_solver = None
         
         def _process(instance):
             solver = deepcopy(self)
             results = solver.solve(instance)
-            solver._clear()
+            solver.internal_solver = None
+            if not collect_training_data:
+                solver.components = {}
             return solver, results
 
         solver_result_pairs = Parallel(n_jobs=n_jobs)(