This PR addresses the issue of the ExperimentalEmulator being retrained on reinitialization. The ExperimentalEmulator class itself does not retrain the model during initialization. However, it was found that in some parts of the codebase, the emulator was being reinitialized and retrained in a loop, which was unnecessary and could lead to performance issues.
Changes made:
Checked the usage of ExperimentalEmulator in summit/tests/test_runner.py and ensured it's not being retrained unnecessarily.
Checked the usage of ExperimentalEmulator in experiments/snar_benchmark/test_snar_experiment.py and ensured it's not being retrained unnecessarily.
Checked the usage of ExperimentalEmulator in experiments/cn_benchmark/test_cn_experiment_MO.py and ensured it's not being retrained unnecessarily.
These changes should prevent the unnecessary retraining of the ExperimentalEmulator and improve the performance of the code.
This PR addresses the issue of the
ExperimentalEmulator
being retrained on reinitialization. TheExperimentalEmulator
class itself does not retrain the model during initialization. However, it was found that in some parts of the codebase, the emulator was being reinitialized and retrained in a loop, which was unnecessary and could lead to performance issues.Changes made:
ExperimentalEmulator
insummit/tests/test_runner.py
and ensured it's not being retrained unnecessarily.ExperimentalEmulator
inexperiments/snar_benchmark/test_snar_experiment.py
and ensured it's not being retrained unnecessarily.ExperimentalEmulator
inexperiments/cn_benchmark/test_cn_experiment_MO.py
and ensured it's not being retrained unnecessarily.These changes should prevent the unnecessary retraining of the
ExperimentalEmulator
and improve the performance of the code.Fixes #253.