共用方式為


DatacacheConsumptionConfig 類別

注意

這是實驗性類別,隨時可能會變更。 請參閱 https://aka.ms/azuremlexperimental 以取得詳細資訊。

表示如何將 datacache 傳遞給計算目標。

表示如何將 datacache 傳遞給計算目標。

繼承
builtins.object
DatacacheConsumptionConfig

建構函式

DatacacheConsumptionConfig(datacache_store, dataset, mode='mount', replica_count=None, path_on_compute=None, **kwargs)

參數

datacache_store
DatacacheStore
必要

要用來快取資料集的 datacachestore。

dataset
FileDataset
必要

需要快取的資料集。

mode
str
必要

定義如何將 datacache 傳遞至計算目標。 目前僅支援掛接模式。

replica_count
<xref:Int>, Optional, <xref:Defaults to 1>
必要

作業要使用的複本數目。

path_on_compute
str, Optional
必要

計算上要讓資料可供使用的目標路徑。

datacache_store
DatacacheStore
必要

要用來快取資料集的 datacachestore。

dataset
FileDataset
必要

需要快取的資料集。

mode
str
必要

定義如何將 datacache 傳遞至計算目標。 目前僅支援掛接模式。

replica_count
<xref:Int>, Optional
必要

定型作業要使用的複本數目。

path_on_compute
str, Optional
必要

計算上要讓資料可供使用的目標路徑。

方法

as_mount

設定要從要求複本數目掛接的 datacache 模式。

在提交的執行中,必要的複本會掛接至計算目標上的本機路徑。 計算上的本機路徑可以從執行中的引數擷取。

as_mount

設定要從要求複本數目掛接的 datacache 模式。

在提交的執行中,必要的複本會掛接至計算目標上的本機路徑。 計算上的本機路徑可以從執行中的引數擷取。

as_mount(path_on_compute=None, replica_count=None)

參數

path_on_compute
str, Optional
預設值: None

計算上要讓資料可供使用的目標路徑。

replica_count
<xref:Int>
預設值: None

複本數目