Source code for otx.core.utils.utils
# Copyright (c) 2024 Intel Corporation
# SPDX-License-Identifier: Apache-2.0
#
"""Utility functions."""
from __future__ import annotations
import importlib
from collections import defaultdict
from multiprocessing import cpu_count
from typing import TYPE_CHECKING, Any
import torch
from datumaro.components.annotation import AnnotationType, LabelCategories
if TYPE_CHECKING:
from datumaro import Dataset as DmDataset
[docs]
def is_ckpt_from_otx_v1(ckpt: dict) -> bool:
"""Check the checkpoint where it comes from.
Args:
ckpt (dict): the checkpoint file
Returns:
bool: True means the checkpoint comes from otx1
"""
return "model" in ckpt and ckpt["VERSION"] == 1
[docs]
def is_ckpt_for_finetuning(ckpt: dict) -> bool:
"""Check the checkpoint will be used to finetune.
Args:
ckpt (dict): the checkpoint file
Returns:
bool: True means the checkpoint will be used to finetune.
"""
return "state_dict" in ckpt
[docs]
def get_adaptive_num_workers(num_dataloader: int = 1) -> int | None:
"""Measure appropriate num_workers value and return it."""
num_gpus = torch.cuda.device_count()
if num_gpus == 0:
return None
return min(cpu_count() // (num_dataloader * num_gpus), 8) # max available num_workers is 8
[docs]
def get_idx_list_per_classes(dm_dataset: DmDataset, use_string_label: bool = False) -> dict[int | str, list[int]]:
"""Compute class statistics."""
stats: dict[int | str, list[int]] = defaultdict(list)
labels = dm_dataset.categories().get(AnnotationType.label, LabelCategories())
for item_idx, item in enumerate(dm_dataset):
for ann in item.annotations:
if use_string_label:
stats[labels.items[ann.label].name].append(item_idx)
else:
stats[ann.label].append(item_idx)
# Remove duplicates in label stats idx: O(n)
for k in stats:
stats[k] = list(dict.fromkeys(stats[k]))
return stats
[docs]
def import_object_from_module(obj_path: str) -> Any: # noqa: ANN401
"""Get object from import format string."""
module_name, obj_name = obj_path.rsplit(".", 1)
module = importlib.import_module(module_name)
return getattr(module, obj_name)
[docs]
def remove_state_dict_prefix(state_dict: dict[str, Any], prefix: str) -> dict[str, Any]:
"""Remove prefix from state_dict keys."""
new_state_dict = {}
for key, value in state_dict.items():
new_key = key.replace(prefix, "")
new_state_dict[new_key] = value
return new_state_dict