dragon-ml-toolbox 20.9.0__tar.gz → 20.11.0__tar.gz

{dragon_ml_toolbox-20.9.0/dragon_ml_toolbox.egg-info → dragon_ml_toolbox-20.11.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dragon-ml-toolbox
-Version: 20.9.0
+Version: 20.11.0
 Summary: Complete pipelines and helper tools for data science and machine learning projects.
 Author-email: Karl Luigi Loza Vidaurre <luigiloza@gmail.com>
 License-Expression: MIT

{dragon_ml_toolbox-20.9.0 → dragon_ml_toolbox-20.11.0/dragon_ml_toolbox.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dragon-ml-toolbox
-Version: 20.9.0
+Version: 20.11.0
 Summary: Complete pipelines and helper tools for data science and machine learning projects.
 Author-email: Karl Luigi Loza Vidaurre <luigiloza@gmail.com>
 License-Expression: MIT

{dragon_ml_toolbox-20.9.0 → dragon_ml_toolbox-20.11.0}/ml_tools/ML_callbacks/_checkpoint.py

@@ -79,7 +79,7 @@ class DragonModelCheckpoint(_Callback):
 
     def on_train_begin(self, logs=None):
         """Reset file tracking state when training starts.
-        NOTE: Do nOT reset self.best here if it differs from the default. This allows the Trainer to restore 'best' from a checkpoint before calling train()."""
+        NOTE: Do not reset self.best here if it differs from the default. This allows the Trainer to restore 'best' from a checkpoint before calling train()."""
         self.best_checkpoints = []
         self.recent_checkpoints = []
         

{dragon_ml_toolbox-20.9.0 → dragon_ml_toolbox-20.11.0}/ml_tools/ML_evaluation_captum/_ML_evaluation_captum.py

@@ -126,7 +126,8 @@ def captum_feature_importance(model: nn.Module,
             feature_names=feature_names,
             save_dir=save_dir_path,
             n_steps=n_steps,
-            file_suffix=f"_{clean_name}"
+            file_suffix=f"_{clean_name}",
+            target_name=name  # Pass original name for plotting
         )
 
 
@@ -137,7 +138,8 @@ def _process_single_target(ig: 'IntegratedGradients', # type: ignore
                            feature_names: Optional[list[str]],
                            save_dir: Path,
                            n_steps: int,
-                           file_suffix: str):
+                           file_suffix: str,
+                           target_name: str = ""):
     """
     Private helper to run the attribution, aggregation, and saving for a single context.
     """
@@ -214,7 +216,11 @@ def _process_single_target(ig: 'IntegratedGradients', # type: ignore
     plt.xlabel("Mean Absolute Attribution")
     
     title = "Feature Importance"
-    if file_suffix:
+    
+    # Use the original target name if provided, otherwise fallback to suffix logic
+    if target_name:
+        title += f" ({target_name})"
+    elif file_suffix:
         # Remove the leading underscore for the title
         clean_suffix = file_suffix.lstrip("_").replace("_", " ")
         title += f" ({clean_suffix})"
@@ -228,7 +234,9 @@ def _process_single_target(ig: 'IntegratedGradients', # type: ignore
     plt.savefig(plot_path)
     plt.close()
     
-    _LOGGER.info(f"🔬 Captum explanation for target '{file_suffix.lstrip("_").replace("_", " ")}' saved to '{save_dir.name}'")
+    # Use target_name for logging if available, otherwise fallback to cleaning the suffix
+    log_name = target_name if target_name else file_suffix.lstrip("_").replace("_", " ")
+    _LOGGER.info(f"🔬 Captum explanation for target '{log_name}' saved to '{save_dir.name}'")
 
 
 def captum_image_heatmap(model: nn.Module,

{dragon_ml_toolbox-20.9.0 → dragon_ml_toolbox-20.11.0}/ml_tools/ML_trainer/_base_trainer.py

@@ -276,6 +276,16 @@ class _BaseDragonTrainer(ABC):
         except Exception as e:
             _LOGGER.error(f"Failed to load checkpoint from '{p}': {e}")
             raise
+    
+    def load_checkpoint(self, path: Union[str, Path], verbose: int = 3):
+        """
+        Loads a specific checkpoint state into the model, optimizer, and scheduler.
+
+        Args:
+            path (str | Path): Path to the .pth checkpoint file.
+            verbose (int): Verbosity level for logging.
+        """
+        self._load_checkpoint(path=path, verbose=verbose)
 
     def fit(self, 
             save_dir: Union[str,Path],
@@ -366,7 +376,7 @@ class _BaseDragonTrainer(ABC):
         self.device = self._validate_device(device)
         self.model.to(self.device)
         _LOGGER.info(f"Trainer and model moved to {self.device}.")
-        
+    
     def _load_model_state_wrapper(self, model_checkpoint: Union[Path, Literal['best', 'current']], verbose: int = 2):
         """
         Private helper to load the correct model state_dict based on user's choice.

{dragon_ml_toolbox-20.9.0 → dragon_ml_toolbox-20.11.0}/ml_tools/keys/_keys.py

@@ -113,7 +113,7 @@ class PyTorchCheckpointKeys:
     EPOCH = "epoch"
     BEST_SCORE = "best_score"
     HISTORY = "history"
-    CHECKPOINT_NAME = "PyModelCheckpoint"
+    CHECKPOINT_NAME = "DragonCheckpoint"
     
     ### Finalized config
     # EPOCH
@@ -296,8 +296,8 @@ class _EvaluationConfig:
     DPI = 400
     LABEL_PADDING = 10
     # large sizes for SVG layout to accommodate large fonts
-    REGRESSION_PLOT_SIZE = (12, 8)
-    SEQUENCE_PLOT_SIZE = (12, 8)
+    REGRESSION_PLOT_SIZE = (10, 7)
+    SEQUENCE_PLOT_SIZE = (10, 7)
     CLASSIFICATION_PLOT_SIZE = (9, 9)
     # Loss plot
     LOSS_PLOT_SIZE = (18, 9)

{dragon_ml_toolbox-20.9.0 → dragon_ml_toolbox-20.11.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "dragon-ml-toolbox"
-version = "20.9.0"
+version = "20.11.0"
 description = "Complete pipelines and helper tools for data science and machine learning projects."
 authors = [
     { name = "Karl Luigi Loza Vidaurre", email = "luigiloza@gmail.com" }