update integration after recent deprecations

pyproject.toml

@@ -14,7 +14,7 @@ python = "^3.7"
 # Python lack of functionalities from future versions
 importlib-metadata = { version = "*", python = "<3.8" }
 
-neptune-client = ">=0.10.0"
+neptune-client = ">=0.16.16"
 numpy = "<1.24.0"
 fvcore = "<0.1.5.post20221220"
 

src/neptune_detectron2/impl/__init__.py

@@ -18,6 +18,8 @@
 
 The class is used for automatic metadata logging to Neptune, during training and validation of detectron2 models."""
 
+from __future__ import annotations
+
 __all__ = [
     "__version__",
     "NeptuneHook",
@@ -43,8 +45,10 @@
 import detectron2
 from detectron2.checkpoint import Checkpointer
 from detectron2.engine import hooks
+from neptune.new.handler import Handler
 from neptune.new.metadata_containers import Run
 from neptune.new.types import File
+from neptune.new.utils import stringify_unsupported
 from torch.nn import Module
 
 from neptune_detectron2.impl.version import __version__
@@ -56,7 +60,7 @@ class NeptuneHook(hooks.HookBase):
     """Hook implementation that sends the logs to Neptune.
 
     Args:
-        run: Pass a Neptune run object if you want to continue logging to an existing run.
+        run: Pass a Neptune run or handler object if you want to continue logging to an existing run.
             Learn more about resuming runs in the docs: https://docs.neptune.ai/logging/to_existing_object
         base_namespace: In the Neptune run, the root namespace that will contain all the logged metadata.
         smoothing_window_size: How often NeptuneHook should log metrics (and checkpoints, if
@@ -90,7 +94,7 @@ class NeptuneHook(hooks.HookBase):
     def __init__(
         self,
         *,
-        run: Optional[Run] = None,
+        run: Optional[Run | Handler] = None,
         base_namespace: str = "training",
         smoothing_window_size: int = 20,
         log_model: bool = False,
@@ -103,14 +107,14 @@ def __init__(
 
         self._verify_window_size()
 
-        self._run = neptune.init_run(**kwargs) if not isinstance(run, Run) else run
+        self._run = neptune.init_run(**kwargs) if not run else run
 
-        verify_type("run", self._run, Run)
+        verify_type("run", self._run, (Run, Handler))
 
         if base_namespace.endswith("/"):
             self._base_namespace = base_namespace[:-1]
 
-        self.base_handler = self._run[base_namespace]
+        self.base_handler = self._run[self._base_namespace]
 
     def _verify_window_size(self) -> None:
         if self._window_size <= 0:
@@ -123,7 +127,7 @@ def _log_integration_version(self) -> None:
 
     def _log_config(self) -> None:
         if hasattr(self.trainer, "cfg") and isinstance(self.trainer.cfg, dict):
-            self.base_handler["config"] = self.trainer.cfg
+            self.base_handler["config"] = stringify_unsupported(self.trainer.cfg)
 
     def _log_model(self) -> None:
         if hasattr(self.trainer, "model") and isinstance(self.trainer.model, Module):
@@ -148,7 +152,7 @@ def _log_checkpoint(self, final: bool = False) -> None:
     def _log_metrics(self) -> None:
         storage = detectron2.utils.events.get_event_storage()
         for k, (v, _) in storage.latest_with_smoothing_hint(self._window_size).items():
-            self.base_handler[f"metrics/{k}"].log(v)
+            self.base_handler[f"metrics/{k}"].append(v)
 
     def _can_save_checkpoint(self) -> bool:
         return hasattr(self.trainer, "checkpointer") and isinstance(self.trainer.checkpointer, Checkpointer)
@@ -177,5 +181,8 @@ def after_train(self) -> None:
         if self.log_model:
             self._log_checkpoint(final=True)
 
+        if isinstance(self._run, Handler):
+            self._run = self._run.get_root_object()
+
         self._run.sync()
         self._run.stop()