Change _optimizer to _optimizer_func

spotlight/factorization/explicit.py

@@ -47,7 +47,7 @@ class ExplicitFactorizationModel(object):
         L2 loss penalty.
     learning_rate: float, optional
         Initial learning rate.
-    optimizer: function, optional
+    optimizer_func: function, optional
         Function that takes in module parameters as the first argument and
         returns an instance of a Pytorch optimizer. Overrides l2 and learning
         rate if supplied. If no optimizer supplied, then use ADAM by default.
@@ -66,7 +66,7 @@ def __init__(self,
                  batch_size=256,
                  l2=0.0,
                  learning_rate=1e-2,
-                 optimizer=None,
+                 optimizer_func=None,
                  use_cuda=False,
                  sparse=False,
                  random_state=None):
@@ -82,12 +82,13 @@ def __init__(self,
         self._l2 = l2
         self._use_cuda = use_cuda
         self._sparse = sparse
-        self._optimizer = optimizer
+        self._optimizer_func = optimizer_func
         self._random_state = random_state or np.random.RandomState()
 
         self._num_users = None
         self._num_items = None
         self._net = None
+        self._optimizer = None
 
         set_seed(self._random_state.randint(-10**8, 10**8),
                  cuda=self._use_cuda)
@@ -122,14 +123,14 @@ def fit(self, interactions, verbose=False):
             self._use_cuda
         )
 
-        if self._optimizer is None:
+        if self._optimizer_func is None:
             self._optimizer = optim.Adam(
                 self._net.parameters(),
                 weight_decay=self._l2,
                 lr=self._learning_rate
             )
         else:
-            self._optimizer = self._optimizer(self._net.parameters())
+            self._optimizer = self._optimizer_func(self._net.parameters())
 
         if self._loss == 'regression':
             loss_fnc = regression_loss

spotlight/factorization/implicit.py

@@ -55,7 +55,7 @@ class ImplicitFactorizationModel(object):
         L2 loss penalty.
     learning_rate: float, optional
         Initial learning rate.
-    optimizer: function, optional
+    optimizer_func: function, optional
         Function that takes in module parameters as the first argument and
         returns an instance of a Pytorch optimizer. Overrides l2 and learning
         rate if supplied. If no optimizer supplied, then use ADAM by default.
@@ -74,7 +74,7 @@ def __init__(self,
                  batch_size=256,
                  l2=0.0,
                  learning_rate=1e-2,
-                 optimizer=None,
+                 optimizer_func=None,
                  use_cuda=False,
                  sparse=False,
                  random_state=None):
@@ -92,12 +92,13 @@ def __init__(self,
         self._l2 = l2
         self._use_cuda = use_cuda
         self._sparse = sparse
-        self._optimizer = optimizer
+        self._optimizer_func = optimizer_func
         self._random_state = random_state or np.random.RandomState()
 
         self._num_users = None
         self._num_items = None
         self._net = None
+        self._optimizer = None
 
         set_seed(self._random_state.randint(-10**8, 10**8),
                  cuda=self._use_cuda)
@@ -139,7 +140,7 @@ def fit(self, interactions, verbose=False):
                 lr=self._learning_rate
             )
         else:
-            self._optimizer = self._optimizer(self._net.parameters())
+            self._optimizer = self._optimizer_func(self._net.parameters())
 
         if self._loss == 'pointwise':
             loss_fnc = pointwise_loss

spotlight/sequence/implicit.py

@@ -47,7 +47,7 @@ class ImplicitSequenceModel(object):
         L2 loss penalty.
     learning_rate: float, optional
         Initial learning rate.
-    optimizer: function, optional
+    optimizer_func: function, optional
         Function that takes in module parameters as the first argument and
         returns an instance of a Pytorch optimizer. Overrides l2 and learning
         rate if supplied. If no optimizer supplied, then use ADAM by default.
@@ -87,7 +87,7 @@ def __init__(self,
                  batch_size=256,
                  l2=0.0,
                  learning_rate=1e-2,
-                 optimizer=None,
+                 optimizer_func=None,
                  use_cuda=False,
                  sparse=False,
                  random_state=None):
@@ -111,11 +111,12 @@ def __init__(self,
         self._l2 = l2
         self._use_cuda = use_cuda
         self._sparse = sparse
-        self._optimizer = optimizer
+        self._optimizer_func = optimizer_func
         self._random_state = random_state or np.random.RandomState()
 
         self._num_items = None
         self._net = None
+        self._optimizer = None
 
         set_seed(self._random_state.randint(-10**8, 10**8),
                  cuda=self._use_cuda)
@@ -163,7 +164,7 @@ def fit(self, interactions, verbose=False):
                 lr=self._learning_rate
             )
         else:
-            self._optimizer = self._optimizer(self._net.parameters())
+            self._optimizer = self._optimizer_func(self._net.parameters())
 
         if self._loss == 'pointwise':
             loss_fnc = pointwise_loss

tests/factorization/test_implicit.py

@@ -66,7 +66,7 @@ def adagrad_optimizer(model_params,
     model = ImplicitFactorizationModel(loss='bpr',
                                        n_iter=10,
                                        batch_size=1024,
-                                       optimizer=adagrad_optimizer)
+                                       optimizer_func=adagrad_optimizer)
     model.fit(train)
 
     mrr = mrr_score(model, test, train=train).mean()