first version (#1494)

src/super_gradients/common/environment/package_utils.py

@@ -0,0 +1,7 @@
+import pkg_resources
+from typing import Dict
+
+
+def get_installed_packages() -> Dict[str, str]:
+    """Map all the installed packages to their version."""
+    return {package.key.lower(): package.version for package in pkg_resources.working_set}

src/super_gradients/sanity_check/env_sanity_check.py

@@ -8,6 +8,7 @@
 
 from super_gradients.common.abstractions.abstract_logger import get_logger
 from super_gradients.common.environment.ddp_utils import is_main_process
+from super_gradients.common.environment.package_utils import get_installed_packages
 
 logger = get_logger(__name__, "DEBUG")
 
@@ -79,7 +80,7 @@ def check_packages():
     """
     test_name = "installed packages"
 
-    installed_packages = {package.key.lower(): package.version for package in pkg_resources.working_set}
+    installed_packages = get_installed_packages()
     requirements = get_requirements(use_pro_requirements="deci-platform-client" in installed_packages)
 
     if requirements is None:

src/super_gradients/training/sg_trainer/sg_trainer.py

@@ -12,7 +12,7 @@
 import torch.nn
 import torchmetrics
 from omegaconf import DictConfig, OmegaConf
-from piptools.scripts.sync import _get_installed_distributions
+
 from torch import nn
 from torch.cuda.amp import GradScaler, autocast
 from torch.utils.data import DataLoader, SequentialSampler
@@ -36,6 +36,7 @@
 from super_gradients.common.factories.list_factory import ListFactory
 from super_gradients.common.factories.losses_factory import LossesFactory
 from super_gradients.common.factories.metrics_factory import MetricsFactory
+from super_gradients.common.environment.package_utils import get_installed_packages
 
 from super_gradients.training import utils as core_utils, models, dataloaders
 from super_gradients.training.datasets.samplers import RepeatAugSampler
@@ -1800,8 +1801,7 @@ def _get_hyper_param_config(self):
         }
         # ADD INSTALLED PACKAGE LIST + THEIR VERSIONS
         if self.training_params.log_installed_packages:
-            pkg_list = list(map(lambda pkg: str(pkg), _get_installed_distributions()))
-            additional_log_items["installed_packages"] = pkg_list
+            additional_log_items["installed_packages"] = get_installed_packages()
 
         dataset_params = {
             "train_dataset_params": self.train_loader.dataset.dataset_params if hasattr(self.train_loader.dataset, "dataset_params") else None,