Enum PartitionStrategy

Namespace

AiDotNet.Enums

Assembly

AiDotNet.dll

Strategies for partitioning models between cloud and edge devices.

public enum PartitionStrategy

Fields

Adaptive = 3

Adaptively determine partition based on runtime conditions (network speed, device power, battery level, model size). Best for dynamic environments.

Balanced = 2

Balanced partition - splits model in the middle. Good general-purpose strategy.

EarlyLayers = 0

Execute early layers on edge, rest on cloud. Good for local preprocessing before cloud processing.

LateLayers = 1

Execute most layers on edge, only final on cloud. Good for powerful edge devices that can handle most processing.

Manual = 4

Manual partition specification - you control exactly where to split. For advanced users with specific requirements.

Remarks

For Beginners: Sometimes you want to split an AI model so that part of it runs on a local device (edge) and part runs in the cloud. This is useful for:

• Reducing bandwidth by processing some data locally
• Improving privacy by keeping sensitive data on the device
• Balancing speed (edge processing) with power (cloud processing)

Different strategies determine where to split the model:

• EarlyLayers: First few layers run on edge, rest in cloud. Good for preprocessing data locally before sending it to the cloud.
• LateLayers: Most processing on edge, only final layers in cloud. Good for devices with decent processing power.
• Balanced: Split in the middle. Good general-purpose strategy.
• Adaptive: Automatically determines the best split based on network speed, device power, and battery level.
• Manual: You specify exactly where to split. For advanced users.