diff --git a/pytorch_lightning/models/trainer.py b/pytorch_lightning/models/trainer.py
index 776f32b1..9af6c394 100644
--- a/pytorch_lightning/models/trainer.py
+++ b/pytorch_lightning/models/trainer.py
@@ -214,7 +214,6 @@ class Trainer(TrainerIO):
         # enable eval mode
         model.zero_grad()
         model.eval()
-        model.from_lightning = True
 
         # disable gradients to save memory
         torch.set_grad_enabled(False)
diff --git a/pytorch_lightning/root_module/root_module.py b/pytorch_lightning/root_module/root_module.py
index 78ef215e..ef43e09a 100644
--- a/pytorch_lightning/root_module/root_module.py
+++ b/pytorch_lightning/root_module/root_module.py
@@ -21,7 +21,6 @@ class LightningModule(GradInformation, ModelIO, OptimizerConfig, ModelHooks):
         self.global_step = 0
         self.loaded_optimizer_states_dict = {}
         self.trainer = None
-        self.from_lightning = True
         self.experiment = None
 
         # track if gpu was requested for checkpointing