tessl/pypi-lightning

The Deep Learning framework to train, deploy, and ship AI products Lightning fast.

—

Pending

Overview

Eval results

Files

Hardware Acceleration

Name: tessl/pypi-lightning
Author: tessl

Support for various hardware accelerators including CPU, CUDA GPUs, Apple Metal Performance Shaders, and Google TPUs with automatic device detection and optimization.

Capabilities

CPU Accelerator

CPU-based training for development, debugging, and CPU-only environments.

class CPUAccelerator:
    def setup_device(self, device: torch.device) -> None:
        """Set up CPU device for training."""

    def get_device_stats(self, device: torch.device) -> Dict[str, Any]:
        """Get CPU device statistics."""

    @staticmethod
    def parse_devices(devices: Union[int, str, List[int]]) -> int:
        """Parse CPU device specification."""

    @staticmethod
    def get_parallel_devices(devices: int) -> List[torch.device]:
        """Get list of CPU devices for parallel training."""

    @staticmethod
    def auto_device_count() -> int:
        """Get number of available CPU cores."""

    @staticmethod
    def is_available() -> bool:
        """Check if CPU is available."""

CUDA Accelerator

NVIDIA GPU acceleration with CUDA support for high-performance training.

class CUDAAccelerator:
    def setup_device(self, device: torch.device) -> None:
        """Set up CUDA device for training."""

    def get_device_stats(self, device: torch.device) -> Dict[str, Any]:
        """Get CUDA device statistics including memory usage."""

    @staticmethod
    def parse_devices(devices: Union[int, str, List[int]]) -> List[int]:
        """Parse CUDA device specification."""

    @staticmethod
    def get_parallel_devices(devices: List[int]) -> List[torch.device]:
        """Get list of CUDA devices for parallel training."""

    @staticmethod
    def auto_device_count() -> int:
        """Get number of available CUDA devices."""

    @staticmethod
    def is_available() -> bool:
        """Check if CUDA is available."""

def find_usable_cuda_devices(num_gpus: int = -1) -> List[int]:
    """
    Find usable CUDA devices.
    
    Args:
        num_gpus: Number of GPUs to find (-1 for all)
        
    Returns:
        List of usable CUDA device IDs
    """

Apple Metal Performance Shaders (MPS)

Apple Silicon GPU acceleration for M1/M2 Macs.

class MPSAccelerator:
    def setup_device(self, device: torch.device) -> None:
        """Set up MPS device for training."""

    def get_device_stats(self, device: torch.device) -> Dict[str, Any]:
        """Get MPS device statistics."""

    @staticmethod
    def parse_devices(devices: Union[int, str, List[int]]) -> int:
        """Parse MPS device specification."""

    @staticmethod
    def get_parallel_devices(devices: int) -> List[torch.device]:
        """Get MPS device for training."""

    @staticmethod
    def auto_device_count() -> int:
        """Get number of available MPS devices."""

    @staticmethod
    def is_available() -> bool:
        """Check if MPS is available."""

XLA Accelerator

Google TPU acceleration using XLA compilation.

class XLAAccelerator:
    def setup_device(self, device: torch.device) -> None:
        """Set up XLA device for training."""

    def get_device_stats(self, device: torch.device) -> Dict[str, Any]:
        """Get XLA device statistics."""

    @staticmethod
    def parse_devices(devices: Union[int, str, List[int]]) -> List[int]:
        """Parse XLA device specification."""

    @staticmethod
    def get_parallel_devices(devices: List[int]) -> List[torch.device]:
        """Get list of XLA devices for parallel training."""

    @staticmethod
    def auto_device_count() -> int:
        """Get number of available XLA devices."""

    @staticmethod
    def is_available() -> bool:
        """Check if XLA is available."""

Install with Tessl CLI