Fix Mixing hparams and arguments in LightningModule (#1505)

* Attempt to fix #1468

* Remove the if statement, it doesn't actually make any difference

* Update docs

* Correct warnings I caused in the last commit

* Add to changelog

* Actually add to changelog

* Clarify documentation and examples

* Update CHANGELOG.md

Co-authored-by: Jirka Borovec <Borda@users.noreply.github.com>

CHANGELOG.md

@@ -39,6 +39,8 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 - Fixed loggers - flushing last logged metrics even before continue, e.g. `trainer.test()` results ([#1459](https://github.com/PyTorchLightning/pytorch-lightning/pull/1459))
 
+- Fixed LightningModule - Mixing hparams and arguments in `LightningModule.__init__()` crashes load_from_checkpoint() ([#1505](https://github.com/PyTorchLightning/pytorch-lightning/pull/1505))
+
 - Added a missing call to the `on_before_zero_grad` model hook ([#1493](https://github.com/PyTorchLightning/pytorch-lightning/pull/1493)). 
 
 - 

pytorch_lightning/core/lightning.py

@@ -1434,6 +1434,7 @@ def load_from_checkpoint(
         it stores the hyperparameters in the checkpoint if you initialized your :class:`LightningModule`
         with an argument called ``hparams`` which is a :class:`~argparse.Namespace`
         (output of :meth:`~argparse.ArgumentParser.parse_args` when parsing command line arguments).
+        Any other arguments specified through \*args and \*\*kwargs will be passed to the model.
 
         Example:
             .. code-block:: python
@@ -1493,7 +1494,7 @@ def __init__(self, hparams):
                 # or load passing whatever args the model takes to load
                 MyLightningModule.load_from_checkpoint(
                     'path/to/checkpoint.ckpt',
-                    learning_rate=0.1,
+                    learning_rate=0.1, # These arguments will be passed to the model using **kwargs
                     layers=2,
                     pretrained_model=some_model
                 )
@@ -1544,10 +1545,7 @@ def _load_model_state(cls, checkpoint: Dict[str, Any], *args, **kwargs) -> 'Ligh
 
         # load the state_dict on the model automatically
         model_args = [hparams] if hparams else []
-        if len(model_args) > 0:
-            model = cls(*model_args)
-        else:
-            model = cls(*args, **kwargs)
+        model = cls(*model_args, *args, **kwargs)
         model.load_state_dict(checkpoint['state_dict'])
 
         # give model a chance to load something