keystoneml.workflow.AutoCacheRule
The memory budget (bytes)
The scales to sample at (average number of desired data points per partition)
The number of profiling samples to take per scale
The memory budget (bytes)
The number of profiling samples to take per scale
The scales to sample at (average number of desired data points per partition)
Greedy caching strategy given a memory budget.
If no maxMem is provided, the memory budget defaults to 75% of the remaining memory on the cluster (given whatever RDDs are already cached)
The memory budget (bytes)
The scales to sample at (average number of desired data points per partition)
The number of profiling samples to take per scale