Update model_checkpoint.py

remove trainer/lightning modules types to avoid circular import

Update gradient_accumulation_scheduler.py

pytorch_lightning/callbacks/gradient_accumulation_scheduler.py

@@ -56,7 +56,9 @@ def __init__(self, scheduling: dict):
 
         minimal_epoch = min(scheduling.keys())
         if minimal_epoch < 0:
-            raise IndexError(f"Epochs indexing from 1, epoch {minimal_epoch} cannot be interpreted correct")
+            raise IndexError(
+                f"Epochs indexing from 1, epoch {minimal_epoch} cannot be interpreted correct"
+            )
         if minimal_epoch != 0:  # if user didnt define first epoch accumulation factor
             scheduling.update({0: 1})
 

pytorch_lightning/callbacks/model_checkpoint.py

@@ -416,7 +416,16 @@ def on_validation_end(self, trainer, pl_module):
                 self._del_model(self.last_model_path)
 
     def _do_check_save(
+<<<<<<< HEAD
         self, filepath: str, current: torch.Tensor, epoch: int, trainer, pl_module
+=======
+        self,
+        filepath: str,
+        current: torch.Tensor,
+        epoch: int,
+        trainer,
+        pl_module,
+>>>>>>> Update model_checkpoint.py
     ):
         # remove kth
 
@@ -452,7 +461,13 @@ def _do_check_save(
                 if cur_path != filepath:
                     self._del_model(cur_path)
 
+<<<<<<< HEAD
     def on_save_checkpoint(self, trainer, pl_module) -> Dict[str, Any]:
+=======
+    def on_save_checkpoint(
+        self, trainer, pl_module
+    ) -> Dict[str, Any]:
+>>>>>>> Update model_checkpoint.py
         return {
             "best_model_score": self.best_model_score,
             "best_model_path": self.best_model_path,