nshtrainer 1.4.0__tar.gz → 1.5.0__tar.gz

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: nshtrainer
-Version: 1.4.0
+Version: 1.5.0
 Summary: 
 Author: Nima Shoghi
 Author-email: nimashoghi@gmail.com
@@ -14,7 +14,7 @@ Provides-Extra: extra
 Requires-Dist: GitPython ; extra == "extra"
 Requires-Dist: huggingface-hub ; extra == "extra"
 Requires-Dist: lightning
-Requires-Dist: nshconfig (>0.39)
+Requires-Dist: nshconfig (>=0.43)
 Requires-Dist: nshrunner ; extra == "extra"
 Requires-Dist: nshutils ; extra == "extra"
 Requires-Dist: numpy

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/pyproject.toml

@@ -1,13 +1,13 @@
 [project]
 name = "nshtrainer"
-version = "1.4.0"
+version = "1.5.0"
 description = ""
 authors = [{ name = "Nima Shoghi", email = "nimashoghi@gmail.com" }]
 requires-python = ">=3.10,<4.0"
 readme = "README.md"
 
 dependencies = [
-    "nshconfig>0.39",
+    "nshconfig>=0.43",
     "psutil",
     "numpy",
     "torch",
@@ -47,7 +47,12 @@ deprecateTypingAliases = true
 strictListInference = true
 strictDictionaryInference = true
 strictSetInference = true
-reportPrivateImportUsage = false
+reportPrivateImportUsage = "none"
+reportMatchNotExhaustive = "warning"
+reportOverlappingOverload = "warning"
+reportUnnecessaryTypeIgnoreComment = "warning"
+reportImplicitOverride = "warning"
+reportIncompatibleMethodOverride = "information"
 
 [tool.ruff.lint]
 select = ["FA102", "FA100"]

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/_callback.py

@@ -8,38 +8,46 @@ from lightning.pytorch import LightningModule
 from lightning.pytorch.callbacks import Callback as _LightningCallback
 from lightning.pytorch.utilities.types import STEP_OUTPUT
 from torch.optim import Optimizer
+from typing_extensions import override
 
 if TYPE_CHECKING:
     from .trainer import Trainer
 
 
 class NTCallbackBase(_LightningCallback):
+    @override
     def setup(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule, stage: str
     ) -> None:
         """Called when fit, validate, test, predict, or tune begins."""
 
+    @override
     def teardown(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule, stage: str
     ) -> None:
         """Called when fit, validate, test, predict, or tune ends."""
 
+    @override
     def on_fit_start(self, trainer: Trainer, pl_module: LightningModule) -> None:  # pyright: ignore[reportIncompatibleMethodOverride]
         """Called when fit begins."""
 
+    @override
     def on_fit_end(self, trainer: Trainer, pl_module: LightningModule) -> None:  # pyright: ignore[reportIncompatibleMethodOverride]
         """Called when fit ends."""
 
+    @override
     def on_sanity_check_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the validation sanity check starts."""
 
+    @override
     def on_sanity_check_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the validation sanity check ends."""
 
+    @override
     def on_train_batch_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -49,6 +57,7 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called when the train batch begins."""
 
+    @override
     def on_train_batch_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -65,11 +74,13 @@ class NTCallbackBase(_LightningCallback):
 
         """
 
+    @override
     def on_train_epoch_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the train epoch begins."""
 
+    @override
     def on_train_epoch_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
@@ -81,10 +92,12 @@ class NTCallbackBase(_LightningCallback):
         .. code-block:: python
 
             class MyLightningModule(L.LightningModule):
+                 @override
                  def __init__(self):
                     super().__init__() # pyright: ignore[reportIncompatibleMethodOverride]
                     self.training_step_outputs = []
 
+                 @override
                  def training_step(self):
                     loss = ... # pyright: ignore[reportIncompatibleMethodOverride]
                     self.training_step_outputs.append(loss)
@@ -92,6 +105,7 @@ class NTCallbackBase(_LightningCallback):
 
 
             class MyCallback(L.Callback):
+                 @override
                  def on_train_epoch_end(self, trainer, pl_module):
                     # do something with all training_step outputs, for example: # pyright: ignore[reportIncompatibleMethodOverride]
                     epoch_mean = torch.stack(pl_module.training_step_outputs).mean()
@@ -101,36 +115,43 @@ class NTCallbackBase(_LightningCallback):
 
         """
 
+    @override
     def on_validation_epoch_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the val epoch begins."""
 
+    @override
     def on_validation_epoch_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the val epoch ends."""
 
+    @override
     def on_test_epoch_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the test epoch begins."""
 
+    @override
     def on_test_epoch_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the test epoch ends."""
 
+    @override
     def on_predict_epoch_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the predict epoch begins."""
 
+    @override
     def on_predict_epoch_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the predict epoch ends."""
 
+    @override
     def on_validation_batch_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -141,6 +162,7 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called when the validation batch begins."""
 
+    @override
     def on_validation_batch_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -152,6 +174,7 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called when the validation batch ends."""
 
+    @override
     def on_test_batch_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -162,6 +185,7 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called when the test batch begins."""
 
+    @override
     def on_test_batch_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -173,6 +197,7 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called when the test batch ends."""
 
+    @override
     def on_predict_batch_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -183,6 +208,7 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called when the predict batch begins."""
 
+    @override
     def on_predict_batch_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -194,36 +220,45 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called when the predict batch ends."""
 
+    @override
     def on_train_start(self, trainer: Trainer, pl_module: LightningModule) -> None:  # pyright: ignore[reportIncompatibleMethodOverride]
         """Called when the train begins."""
 
+    @override
     def on_train_end(self, trainer: Trainer, pl_module: LightningModule) -> None:  # pyright: ignore[reportIncompatibleMethodOverride]
         """Called when the train ends."""
 
+    @override
     def on_validation_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the validation loop begins."""
 
+    @override
     def on_validation_end(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the validation loop ends."""
 
+    @override
     def on_test_start(self, trainer: Trainer, pl_module: LightningModule) -> None:  # pyright: ignore[reportIncompatibleMethodOverride]
         """Called when the test begins."""
 
+    @override
     def on_test_end(self, trainer: Trainer, pl_module: LightningModule) -> None:  # pyright: ignore[reportIncompatibleMethodOverride]
         """Called when the test ends."""
 
+    @override
     def on_predict_start(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called when the predict begins."""
 
+    @override
     def on_predict_end(self, trainer: Trainer, pl_module: LightningModule) -> None:  # pyright: ignore[reportIncompatibleMethodOverride]
         """Called when predict ends."""
 
+    @override
     def on_exception(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -232,7 +267,8 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called when any trainer execution is interrupted by an exception."""
 
-    def state_dict(self) -> dict[str, Any]:  # pyright: ignore[reportIncompatibleMethodOverride]
+    @override
+    def state_dict(self) -> dict[str, Any]:
         """Called when saving a checkpoint, implement to generate callback's ``state_dict``.
 
         Returns:
@@ -241,7 +277,8 @@ class NTCallbackBase(_LightningCallback):
         """
         return {}
 
-    def load_state_dict(self, state_dict: dict[str, Any]) -> None:  # pyright: ignore[reportIncompatibleMethodOverride]
+    @override
+    def load_state_dict(self, state_dict: dict[str, Any]) -> None:
         """Called when loading a checkpoint, implement to reload callback state given callback's ``state_dict``.
 
         Args:
@@ -250,6 +287,7 @@ class NTCallbackBase(_LightningCallback):
         """
         pass
 
+    @override
     def on_save_checkpoint(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -265,6 +303,7 @@ class NTCallbackBase(_LightningCallback):
 
         """
 
+    @override
     def on_load_checkpoint(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -280,16 +319,19 @@ class NTCallbackBase(_LightningCallback):
 
         """
 
+    @override
     def on_before_backward(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule, loss: torch.Tensor
     ) -> None:
         """Called before ``loss.backward()``."""
 
+    @override
     def on_after_backward(  # pyright: ignore[reportIncompatibleMethodOverride]
         self, trainer: Trainer, pl_module: LightningModule
     ) -> None:
         """Called after ``loss.backward()`` and before optimizers are stepped."""
 
+    @override
     def on_before_optimizer_step(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -298,6 +340,7 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called before ``optimizer.step()``."""
 
+    @override
     def on_before_zero_grad(  # pyright: ignore[reportIncompatibleMethodOverride]
         self,
         trainer: Trainer,
@@ -306,7 +349,10 @@ class NTCallbackBase(_LightningCallback):
     ) -> None:
         """Called before ``optimizer.zero_grad()``."""
 
-    def on_checkpoint_saved(  # pyright: ignore[reportIncompatibleMethodOverride]
+    # =================================================================
+    # Our own new callbacks
+    # =================================================================
+    def on_checkpoint_saved(
         self,
         ckpt_path: Path,
         metadata_path: Path | None,
@@ -317,6 +363,7 @@ class NTCallbackBase(_LightningCallback):
         pass
 
 
+@override
 def _call_on_checkpoint_saved(
     trainer: Trainer,
     ckpt_path: str | Path,

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/callbacks/__init__.py

@@ -75,5 +75,5 @@ from .wandb_watch import WandbWatchCallbackConfig as WandbWatchCallbackConfig
 
 CallbackConfig = TypeAliasType(
     "CallbackConfig",
-    Annotated[CallbackConfigBase, callback_registry.DynamicResolution()],
+    Annotated[CallbackConfigBase, callback_registry],
 )

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/callbacks/checkpoint/_base.py

@@ -5,13 +5,13 @@ import string
 from abc import ABC, abstractmethod
 from collections.abc import Callable
 from pathlib import Path
-from typing import TYPE_CHECKING, Any, Generic, Literal, TypeVar
+from typing import TYPE_CHECKING, Any, Generic, Literal
 
 import numpy as np
 import torch
 from lightning.pytorch import Trainer
 from lightning.pytorch.callbacks import Checkpoint
-from typing_extensions import override
+from typing_extensions import TypeVar, override
 
 from ..._checkpoint.metadata import CheckpointMetadata, _generate_checkpoint_metadata
 from ..._checkpoint.saver import link_checkpoint, remove_checkpoint

nshtrainer-1.5.0/src/nshtrainer/callbacks/log_epoch.py

@@ -0,0 +1,136 @@
+from __future__ import annotations
+
+import logging
+import math
+from typing import Any, Literal
+
+from lightning.pytorch import LightningModule, Trainer
+from lightning.pytorch.callbacks import Callback
+from typing_extensions import final, override
+
+from .base import CallbackConfigBase, callback_registry
+
+log = logging.getLogger(__name__)
+
+
+@final
+@callback_registry.register
+class LogEpochCallbackConfig(CallbackConfigBase):
+    name: Literal["log_epoch"] = "log_epoch"
+
+    metric_name: str = "computed_epoch"
+    """The name of the metric to log the epoch as."""
+
+    train: bool = True
+    """Whether to log the epoch during training."""
+
+    val: bool = True
+    """Whether to log the epoch during validation."""
+
+    test: bool = True
+    """Whether to log the epoch during testing."""
+
+    @override
+    def create_callbacks(self, trainer_config):
+        yield LogEpochCallback(self)
+
+
+def _worker_fn(
+    trainer: Trainer,
+    pl_module: LightningModule,
+    num_batches_prop: str,
+    dataloader_idx: int | None = None,
+    *,
+    metric_name: str,
+):
+    if trainer.logger is None:
+        return
+
+    # If trainer.num_{training/val/test}_batches is not set or is nan/inf, we cannot calculate the epoch
+    if not (num_batches := getattr(trainer, num_batches_prop, None)):
+        log.warning(f"Trainer has no valid `{num_batches_prop}`. Cannot log epoch.")
+        return
+
+    # If the trainer has a dataloader_idx, num_batches is a list of num_batches for each dataloader.
+    if dataloader_idx is not None:
+        assert isinstance(num_batches, list), (
+            f"Expected num_batches to be a list, got {type(num_batches)}"
+        )
+        assert 0 <= dataloader_idx < len(num_batches), (
+            f"Expected dataloader_idx to be between 0 and {len(num_batches)}, got {dataloader_idx}"
+        )
+        num_batches = num_batches[dataloader_idx]
+
+    if (
+        not isinstance(num_batches, (int, float))
+        or math.isnan(num_batches)
+        or math.isinf(num_batches)
+    ):
+        log.warning(
+            f"Trainer has no valid `{num_batches_prop}` (got {num_batches=}). Cannot log epoch."
+        )
+        return
+
+    epoch = pl_module.global_step / num_batches
+    pl_module.log(metric_name, epoch, on_step=True, on_epoch=False)
+
+
+class LogEpochCallback(Callback):
+    def __init__(self, config: LogEpochCallbackConfig):
+        super().__init__()
+
+        self.config = config
+
+    @override
+    def on_train_batch_start(
+        self, trainer: Trainer, pl_module: LightningModule, batch: Any, batch_idx: int
+    ):
+        if trainer.logger is None or not self.config.train:
+            return
+
+        _worker_fn(
+            trainer,
+            pl_module,
+            "num_training_batches",
+            metric_name=self.config.metric_name,
+        )
+
+    @override
+    def on_validation_batch_start(
+        self,
+        trainer: Trainer,
+        pl_module: LightningModule,
+        batch: Any,
+        batch_idx: int,
+        dataloader_idx: int = 0,
+    ) -> None:
+        if trainer.logger is None or not self.config.val:
+            return
+
+        _worker_fn(
+            trainer,
+            pl_module,
+            "num_val_batches",
+            dataloader_idx=dataloader_idx,
+            metric_name=self.config.metric_name,
+        )
+
+    @override
+    def on_test_batch_start(
+        self,
+        trainer: Trainer,
+        pl_module: LightningModule,
+        batch: Any,
+        batch_idx: int,
+        dataloader_idx: int = 0,
+    ) -> None:
+        if trainer.logger is None or not self.config.test:
+            return
+
+        _worker_fn(
+            trainer,
+            pl_module,
+            "num_test_batches",
+            dataloader_idx=dataloader_idx,
+            metric_name=self.config.metric_name,
+        )

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/callbacks/print_table.py

@@ -67,14 +67,14 @@ class PrintTableMetricsCallback(Callback):
             }
         self.metrics.append(metrics_dict)
 
-        from rich.console import Console  # type: ignore[reportMissingImports] # noqa
+        from rich.console import Console  # pyright: ignore[reportMissingImports] # noqa
 
         console = Console()
         table = self.create_metrics_table()
         console.print(table)
 
     def create_metrics_table(self):
-        from rich.table import Table  # type: ignore[reportMissingImports] # noqa
+        from rich.table import Table  # pyright: ignore[reportMissingImports] # noqa
 
         table = Table(show_header=True, header_style="bold magenta")
 

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/callbacks/rlp_sanity_checks.py

@@ -38,6 +38,7 @@ class RLPSanityChecksCallbackConfig(CallbackConfigBase):
     def __bool__(self):
         return self.enabled
 
+    @override
     def create_callbacks(self, trainer_config):
         if not self:
             return

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/configs/__init__.py

@@ -111,7 +111,6 @@ from nshtrainer.optimizer import RAdamConfig as RAdamConfig
 from nshtrainer.optimizer import RMSpropConfig as RMSpropConfig
 from nshtrainer.optimizer import RpropConfig as RpropConfig
 from nshtrainer.optimizer import SGDConfig as SGDConfig
-from nshtrainer.optimizer import Union as Union
 from nshtrainer.optimizer import optimizer_registry as optimizer_registry
 from nshtrainer.profiler import AdvancedProfilerConfig as AdvancedProfilerConfig
 from nshtrainer.profiler import BaseProfilerConfig as BaseProfilerConfig
@@ -355,7 +354,6 @@ __all__ = [
     "TorchSyncBatchNormPlugin",
     "TrainerConfig",
     "TransformerEnginePluginConfig",
-    "Union",
     "WandbLoggerConfig",
     "WandbUploadCodeCallbackConfig",
     "WandbWatchCallbackConfig",

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/configs/optimizer/__init__.py

@@ -16,7 +16,6 @@ from nshtrainer.optimizer import RAdamConfig as RAdamConfig
 from nshtrainer.optimizer import RMSpropConfig as RMSpropConfig
 from nshtrainer.optimizer import RpropConfig as RpropConfig
 from nshtrainer.optimizer import SGDConfig as SGDConfig
-from nshtrainer.optimizer import Union as Union
 from nshtrainer.optimizer import optimizer_registry as optimizer_registry
 
 __all__ = [
@@ -34,6 +33,5 @@ __all__ = [
     "RMSpropConfig",
     "RpropConfig",
     "SGDConfig",
-    "Union",
     "optimizer_registry",
 ]

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/loggers/__init__.py

@@ -12,6 +12,5 @@ from .tensorboard import TensorboardLoggerConfig as TensorboardLoggerConfig
 from .wandb import WandbLoggerConfig as WandbLoggerConfig
 
 LoggerConfig = TypeAliasType(
-    "LoggerConfig",
-    Annotated[LoggerConfigBase, logger_registry.DynamicResolution()],
+    "LoggerConfig", Annotated[LoggerConfigBase, logger_registry]
 )

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/loggers/actsave.py

@@ -5,7 +5,7 @@ from typing import Any, Literal
 
 import numpy as np
 from lightning.pytorch.loggers import Logger
-from typing_extensions import final
+from typing_extensions import final, override
 
 from .base import LoggerConfigBase, logger_registry
 
@@ -15,6 +15,7 @@ from .base import LoggerConfigBase, logger_registry
 class ActSaveLoggerConfig(LoggerConfigBase):
     name: Literal["actsave"] = "actsave"
 
+    @override
     def create_logger(self, trainer_config):
         if not self.enabled:
             return None
@@ -24,10 +25,12 @@ class ActSaveLoggerConfig(LoggerConfigBase):
 
 class ActSaveLogger(Logger):
     @property
+    @override
     def name(self):
         return None
 
     @property
+    @override
     def version(self):
         from nshutils import ActSave
 
@@ -37,6 +40,7 @@ class ActSaveLogger(Logger):
         return ActSave._saver._id
 
     @property
+    @override
     def save_dir(self):
         from nshutils import ActSave
 
@@ -45,6 +49,7 @@ class ActSaveLogger(Logger):
 
         return str(ActSave._saver._save_dir)
 
+    @override
     def log_hyperparams(
         self,
         params: dict[str, Any] | Namespace,
@@ -56,6 +61,7 @@ class ActSaveLogger(Logger):
         # Wrap the hparams as a object-dtype np array
         return ActSave.save({"hyperparameters": np.array(params, dtype=object)})
 
+    @override
     def log_metrics(self, metrics: dict[str, float], step: int | None = None) -> None:
         from nshutils import ActSave
 

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/loggers/wandb.py

@@ -63,7 +63,7 @@ class FinishWandbOnTeardownCallback(Callback):
         stage: str,
     ):
         try:
-            import wandb  # type: ignore
+            import wandb
         except ImportError:
             return
 
@@ -139,7 +139,7 @@ class WandbLoggerConfig(CallbackConfigBase, LoggerConfigBase):
         # If `wandb-core` is enabled, we should use the new backend.
         if self.use_wandb_core:
             try:
-                import wandb  # type: ignore
+                import wandb
 
                 # The minimum version that supports the new backend is 0.17.5
                 wandb_version = version.parse(importlib.metadata.version("wandb"))
@@ -151,7 +151,7 @@ class WandbLoggerConfig(CallbackConfigBase, LoggerConfigBase):
                     )
                 # W&B versions 0.18.0 use wandb-core by default
                 elif wandb_version < version.parse("0.18.0"):
-                    wandb.require("core")  # type: ignore
+                    wandb.require("core")
                     log.critical("Using the `wandb-core` backend for WandB.")
             except ImportError:
                 pass
@@ -166,9 +166,9 @@ class WandbLoggerConfig(CallbackConfigBase, LoggerConfigBase):
                     "If you want to use the new `wandb-core` backend, set `use_wandb_core=True`."
                 )
                 try:
-                    import wandb  # type: ignore
+                    import wandb
 
-                    wandb.require("legacy-service")  # type: ignore
+                    wandb.require("legacy-service")
                 except ImportError:
                     pass
 

{nshtrainer-1.4.0 → nshtrainer-1.5.0}/src/nshtrainer/lr_scheduler/base.py

@@ -81,7 +81,7 @@ class LRSchedulerConfigBase(C.Config, ABC):
             scheduler["monitor"] = metadata["monitor"]
         # - `strict`
         if scheduler.get("strict") is None and "strict" in metadata:
-            scheduler["strict"] = metadata["strict"]  # type: ignore
+            scheduler["strict"] = metadata["strict"]
 
         return scheduler