[[["容易理解","easyToUnderstand","thumb-up"],["確實解決了我的問題","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["難以理解","hardToUnderstand","thumb-down"],["資訊或程式碼範例有誤","incorrectInformationOrSampleCode","thumb-down"],["缺少我需要的資訊/範例","missingTheInformationSamplesINeed","thumb-down"],["翻譯問題","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["上次更新時間:2025-07-24 (世界標準時間)。"],[[["This page provides detailed reference information regarding the tuning parameters for the ScaNN Index."],["The `max_num_levels`, `num_leaves`, and `quantizer` parameters are used during index creation and can be adjusted for optimal performance based on the size of the vector dataset."],["The `scann.enable_pca` parameter, enabled by default, uses Principal Component Analysis to reduce embedding size, but can be disabled if recall deteriorates."],["`scann.num_leaves_to_search`, `scann.pre_reordering_num_neighbors`, and `scann.num_search_threads` are query runtime parameters that impact the trade-off between recall and QPS, and can be adjusted based on application requirements."],["Adjusting the `num_leaves` parameter when creating an index can create a more fine-grained index, improving recall and query performance at the cost of longer index creation times."]]],[]]