Fix unfreeze_and_add_param_group expects modules rather than `mod…

…ule` (Lightning-AI#6822)

pytorch_lightning/callbacks/finetuning.py

@@ -77,7 +77,7 @@ def finetune_function(self, pl_module, current_epoch, optimizer, optimizer_idx):
                 # When `current_epoch` is 10, feature_extractor will start training.
                 if current_epoch == self._unfreeze_at_epoch:
                     self.unfreeze_and_add_param_group(
-                        module=pl_module.feature_extractor,
+                        modules=pl_module.feature_extractor,
                         optimizer=optimizer,
                         train_bn=True,
                     )