File size: 2,472 Bytes
452723c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 |
"""Utility functions for torch devices."""
import torch
from src.utils.logging_util import LoggingUtils
logger = LoggingUtils.configure_logger(log_name=__name__)
def get_device(device_str: str):
"""Set device for pytorch operations."""
logger.info(f"PyTorch version: {torch.__version__}")
if "cuda" in device_str and torch.cuda.is_available():
logger.info(f"GPU: {torch.cuda.get_device_name(0)}")
device= torch.device(device_str)
else:
logger.warning("CUDA not available, resorting to CPU.")
device= torch.device("cpu")
logger.info(f"Device: {device}")
return device
def set_all_seed(seed: int = 42, deterministic: bool = True):
"""Set random seed for reproducibility."""
import os
import random
import numpy as np
import torch
if seed == -1:
logger.warning("Set seed disabled.")
else:
# Python built‑ins
random.seed(seed)
os.environ['PYTHONHASHSEED'] = str(seed)
# NumPy
np.random.seed(seed)
# PyTorch (CPU & CUDA)
torch.manual_seed(seed)
if torch.cuda.is_available():
torch.cuda.manual_seed(seed)
torch.cuda.manual_seed_all(seed) # multiple GPUs
# CuDNN backends
if deterministic:
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
def recursively_to_device_float(obj, to_device="cpu"):
"""
Recursively traverse any data structure and convert all torch tensors
to float and optionally move to CPU.
Args:
obj: nested structure (tensor, dict, list, etc.)
to_cpu (bool): whether to move tensors to CPU
"""
if isinstance(to_device, str):
device = torch.device(to_device)
elif isinstance(to_device, torch.device):
device = to_device
else:
raise ValueError(to_device)
if isinstance(obj, torch.Tensor):
return obj.float().to(device)
elif isinstance(obj, dict):
return {k: recursively_to_device_float(v, to_device) for k, v in obj.items()}
elif isinstance(obj, list):
return [recursively_to_device_float(v, to_device) for v in obj]
elif isinstance(obj, tuple):
return tuple(recursively_to_device_float(v, to_device) for v in obj)
elif isinstance(obj, set):
return {recursively_to_device_float(v, to_device) for v in obj}
else:
return obj |