Skip to content
voiage

voiage.backends.gpu_acceleration.GpuAcceleration

Utilities for GPU acceleration and memory management.

Methods

Section titled “Methods”

detect_gpu

Section titled “detect_gpu”
detect_gpu() -> list[jax.Device]

Detect available GPU devices.

Returns: list[jax.Device]

get_memory_info

Section titled “get_memory_info”
get_memory_info() -> dict[str, object]

Get GPU memory information.

Returns: dict[str, object]

optimize_for_gpu

Section titled “optimize_for_gpu”
optimize_for_gpu([positional or keyword] data: object = None) -> object

Optimize data layout for GPU processing.

Parameters:

  • data object

Returns: object

memory_efficient_batch_process

Section titled “memory_efficient_batch_process”
memory_efficient_batch_process([positional or keyword] data_batches: Sequence[np.ndarray] = None, [positional or keyword] process_func: Callable[[list[np.ndarray]], object] = None, [positional or keyword] max_memory_mb: int = 1000) -> object

Process large datasets in memory-efficient batches.

Parameters:

  • data_batches Sequence[np.ndarray]
  • process_func Callable[[list[np.ndarray]], object]
  • max_memory_mb int (default: 1000)

Returns: object