In order to speed up machine learning, I specify my own custom pipelines as follows:
from automatminer import get_preset_config, TPOTAdaptor, MatPipe
config = get_preset_config("express")
config["learner"] = TPOTAdaptor(max_time_mins=6000, n_jobs=36)
But when I use the top command to look for Python process, I find python only use one core when it start "FeatureReducer: Starting fitting." this step, This does not use multiple cores to perform operations like the AutoFeaturizer step. I don't know if it is my incorrect parameter setting or the program itself. I hope my question can be answered, thank you very much!
In addition, if this method cannot make the program parallel and then speed up, I would like to ask if there are other reasonable methods that can be used to speed up machine learning.
In order to speed up machine learning, I specify my own custom pipelines as follows:
from automatminer import get_preset_config, TPOTAdaptor, MatPipe
config = get_preset_config("express")
config["learner"] = TPOTAdaptor(max_time_mins=6000, n_jobs=36)
But when I use the top command to look for Python process, I find python only use one core when it start "FeatureReducer: Starting fitting." this step, This does not use multiple cores to perform operations like the AutoFeaturizer step. I don't know if it is my incorrect parameter setting or the program itself. I hope my question can be answered, thank you very much!
In addition, if this method cannot make the program parallel and then speed up, I would like to ask if there are other reasonable methods that can be used to speed up machine learning.