GpuAccelerationConfig

AI-generated Key Takeaways

• GpuAccelerationConfig is a concrete class representing GPU acceleration configurations and extends AccelerationConfig.

• It includes nested classes for building configurations, selecting the GPU backend, and defining inference priorities and usage preferences.

• The class provides public methods to retrieve details such as the cache directory, quantized inference flag, force backend setting, inference preference, and multiple inference priority levels.

• It also offers a method to retrieve a unique model token string.

• Inherited methods from AccelerationConfig allow retrieving the accelerator name and serializing the configuration.

public class GpuAccelerationConfig extends AccelerationConfig

Concrete class that represents GPU acceleration configs. For more details, see delegate.h

Nested Class Summary

class	GpuAccelerationConfig.Builder		Builder class.
enum	GpuAccelerationConfig.GpuBackend		Which GPU backend to select.
enum	GpuAccelerationConfig.GpuInferencePriority		Relative priorities given by the GPU delegate to different client needs.
enum	GpuAccelerationConfig.GpuInferenceUsage		GPU inference preference for initialization time vs.

Public Method Summary

String	cacheDirectory() Returns serialization cache directory.
boolean	enableQuantizedInference() Returns the enable quantized inference flag.
GpuAccelerationConfig.GpuBackend	forceBackend() Returns the selected GPU backend.
GpuAccelerationConfig.GpuInferenceUsage	inferencePreference() Returns GPU inference preference.
GpuAccelerationConfig.GpuInferencePriority	inferencePriority1() Returns GPU inference priority(1).
GpuAccelerationConfig.GpuInferencePriority	inferencePriority2() Returns GPU inference priority(2).
GpuAccelerationConfig.GpuInferencePriority	inferencePriority3() Returns GPU inference priority(3).
String	modelToken() Returns unique model token string.

Inherited Method Summary

From class com.google.android.gms.tflite.acceleration.AccelerationConfig

String	getAcceleratorName() Returns the accelerator type of this config.
byte[]	serialize() Serializes config as bytes.

From class java.lang.Object

Object	clone()
boolean	equals(Object arg0)
void	finalize()
final Class<?>	getClass()
int	hashCode()
final void	notify()
final void	notifyAll()
String	toString()
final void	wait(long arg0, int arg1)
final void	wait(long arg0)
final void	wait()

Public Methods

public String cacheDirectory ()

Returns serialization cache directory.

public boolean enableQuantizedInference ()

Returns the enable quantized inference flag.

public GpuAccelerationConfig.GpuBackend forceBackend ()

Returns the selected GPU backend.

public GpuAccelerationConfig.GpuInferenceUsage inferencePreference ()

Returns GPU inference preference.

public GpuAccelerationConfig.GpuInferencePriority inferencePriority1 ()

Returns GPU inference priority(1).

public GpuAccelerationConfig.GpuInferencePriority inferencePriority2 ()

Returns GPU inference priority(2).

public GpuAccelerationConfig.GpuInferencePriority inferencePriority3 ()

Returns GPU inference priority(3).

public String modelToken ()

Returns unique model token string.