huawei-noah/vega

searching valid metric is better than re-train valid metric in autoFIS?

开放

#107 创建于 2021年4月8日

 (2 条评论) (0 个反应) (1 位负责人)Python (177 个派生)auto 404
good first issuequestion

仓库指标

星标
 (848 个星标)
PR 合并指标
 (30 天内没有已合并 PR)

描述

Thank you for your excellent codes. I run your autoFIS codes (autoDeepFM), but found that the valid metric after searching (auc=0.7896248002165958) is better than the retrain valid metric (auc=0.7851094769386633) in the log. i.e. [searching valid metric] > [re-train valid metric] > [official valid metric], it looks like the performance improvement is achieved by BN but autofis damages the valid metric. Something wrong ?

the valid metric after searching: image the valid metric after re-train: image

贡献者指南