sklearn_wrapper.py 文件源码

python
阅读 21 收藏 0 点赞 0 评论 0

项目:chainer_sklearn 作者: corochann 项目源码 文件源码
def fit(self, X, y=None, **kwargs):
        """If hyper parameters are set to None, then instance's variable is used,
        this functionality is used Grid search with `set_params` method.
        Also if instance's variable is not set, _default_hyperparam is used. 

        Usage: model.fit(train_dataset) or model.fit(X, y)

        Args:
            train: training dataset, assumes chainer's dataset class 
            test: test dataset for evaluation, assumes chainer's dataset class
            batchsize: batchsize for both training and evaluation
            iterator_class: iterator class used for this training, 
                            currently assumes SerialIterator or MultiProcessIterator
            optimizer: optimizer instance to update parameter
            epoch: training epoch
            out: directory path to save the result
            snapshot_frequency (int): snapshot frequency in epoch. 
                                Negative value indicates not to take snapshot.
            dump_graph: Save computational graph info or not, default is False.
            log_report: Enable LogReport or not
            plot_report: Enable PlotReport or not
            print_report: Enable PrintReport or not
            progress_report: Enable ProgressReport or not
            resume: specify trainer saved path to resume training.

        """
        kwargs = self.filter_sk_params(self.fit_core, kwargs)
        return self.fit_core(X, y, **kwargs)
评论列表
文章目录


问题


面经


文章

微信
公众号

扫码关注公众号