sklearn time cpu time 还是 wall time
Does sklearn time cpu time or wall time
sklearn 的随机和网格搜索实现的返回时间值是 CPU 还是墙上时间?我在文档中只能找到以下内容:
The mean_fit_time, std_fit_time, mean_score_time and std_score_time
are all in seconds.
所以我假设这意味着默认的挂钟时间,这是正确的吗?
经过 source code...
第 748 行:
self.cv_results_ = results
好的,什么是 results
?
第 706 行:
results = self._format_results(
all_candidate_params, scorers, n_splits, all_out)
return results
好的,all_out
是什么?
第 703 行:
all_out.extend(out)
然后out
是:
第 682 行:
out = parallel(delayed(_fit_and_score)(clone(base_estimator),
和 _fit_and_score
是从 sklearn.model_selection._validation.py
导入的,所以我们查看 there...
第 543 行:
fit_time = time.time() - start_time
啊哈!我们开始了。这是挂钟时间,如 Python
的 time.time()
gives you the current Unix time.
sklearn 的随机和网格搜索实现的返回时间值是 CPU 还是墙上时间?我在文档中只能找到以下内容:
The mean_fit_time, std_fit_time, mean_score_time and std_score_time are all in seconds.
所以我假设这意味着默认的挂钟时间,这是正确的吗?
经过 source code...
第 748 行:
self.cv_results_ = results
好的,什么是 results
?
第 706 行:
results = self._format_results(
all_candidate_params, scorers, n_splits, all_out)
return results
好的,all_out
是什么?
第 703 行:
all_out.extend(out)
然后out
是:
第 682 行:
out = parallel(delayed(_fit_and_score)(clone(base_estimator),
和 _fit_and_score
是从 sklearn.model_selection._validation.py
导入的,所以我们查看 there...
第 543 行:
fit_time = time.time() - start_time
啊哈!我们开始了。这是挂钟时间,如 Python
的 time.time()
gives you the current Unix time.