在Google中保存optuna study.pkl

Question

我正在Google Colab上调整我的ML模型，但我不知道如何将该模型保存到pkl。

import time
import optuna

study_name = "/gdrive/MyDrive/Colab Notebooks/test/params_{}".format(time.strftime("%Y%m%d-%H%M%S"))
study=optuna.create_study(study_name, direction='maximize')

代码显示了这个错误：

Could not parse rfc1738 URL from string '/gdrive/MyDrive/Colab Notebooks/test/params_20220217-181559'

我该怎么做才能保存这个模型？

Answer 1

你是说拯救书房？

https://optuna.readthedocs.io/en/stable/faq.html#how-can-i-save-and-resume-studies

我用这个：

install joblib
import joblib

# Let's say I want to save study to savepath + "xgb_optuna_study_batch.pkl"

joblib.dump(study, f"{savepath}xgb_optuna_study_batch.pkl")   # save study

# to load it:
jl = joblib.load(f"{savepath}xgb_optuna_study_batch.pkl")

print(jl.best_trial.params)

# output, for example:
{'lambda': 1.4556073038174557, 'alpha': 0.007250895998233471, 'colsample_bytree': 0.7, 'subsample': 0.8, 'learning_rate': 0.01, 'max_depth': 20, 'random_state': 48, 'min_child_weight': 1}