在知乎上聽說數(shù)據(jù)集放在drive以及content下面的速度會(huì)不同，于是我就嘗試了一下。

?

代碼和數(shù)據(jù)集使用kaggle中的

IEEE Fraud: XGBoost with GPU (Fit in 40s)

數(shù)據(jù)集大小為59w

#############################放在drive下面(記得改路徑)###################################################

CPU times: user 9min 12s, sys: 6min 41s, total: 15min 54s Wall time: 15min 55s XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,colsample_bynode=1, colsample_bytree=0.9, gamma=0,learning_rate=0.05, max_delta_step=0, max_depth=9,min_child_weight=1, missing=-999, n_estimators=500, n_jobs=1,nthread=None, objective='binary:logistic', random_state=2019,reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,silent=None, subsample=0.9, tree_method='gpu_hist', verbosity=1)

#############################放在content下面(記得改路徑)###########################################

CPU times: user 9min 7s, sys: 6min 38s, total: 15min 45s Wall time: 15min 47s XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,colsample_bynode=1, colsample_bytree=0.9, gamma=0,learning_rate=0.05, max_delta_step=0, max_depth=9,min_child_weight=1, missing=-999, n_estimators=500, n_jobs=1,nthread=None, objective='binary:logistic', random_state=2019,reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,silent=None, subsample=0.9, tree_method='gpu_hist', verbosity=1)

總結(jié)

以上是生活随笔為你收集整理的colab把数据放在content下面以及放在drive下面的训练速度比较的全部?jī)?nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。