remove unused parameters in optimizer tests (pytorch#18084)

Summary:
Pull Request resolved: pytorch#18084

data_strategy parameter was not used in some of unit tests for optimizers

Reviewed By: hyuen

Differential Revision: D14487830

fbshipit-source-id: d757cd06aa2965f4c0570a4a18ba090b98820ef4

caffe2/python/operator_test/adadelta_test.py

@@ -144,10 +144,8 @@ def ref_sparse(param, moment, moment_delta, indices, grad, lr, decay,
                              allow_nan=False, allow_infinity=False),
            decay=st.floats(min_value=0.01, max_value=0.99,
                              allow_nan=False, allow_infinity=False),
-           data_strategy=st.data(),
            **hu.gcs)
-    def test_sparse_adadelta_empty(self, inputs, lr, epsilon, decay,
-                                  data_strategy, gc, dc):
+    def test_sparse_adadelta_empty(self, inputs, lr, epsilon, decay, gc, dc):
         param, moment, moment_delta = inputs
         moment = np.abs(moment)
         lr = np.array([lr], dtype=np.float32)

caffe2/python/operator_test/adagrad_test.py

@@ -134,10 +134,9 @@ def test_sparse_adagrad(self, inputs, lr, epsilon, gc, dc):
         epsilon=st.floats(
             min_value=0.01, max_value=0.99, allow_nan=False, allow_infinity=False
         ),
-        data_strategy=st.data(),
         **hu.gcs
     )
-    def test_sparse_adagrad_empty(self, inputs, lr, epsilon, data_strategy, gc, dc):
+    def test_sparse_adagrad_empty(self, inputs, lr, epsilon, gc, dc):
         param, momentum = inputs
         grad = np.empty(shape=(0,) + param.shape[1:], dtype=np.float32)
 
@@ -176,10 +175,9 @@ def test_sparse_adagrad_empty(self, inputs, lr, epsilon, data_strategy, gc, dc):
         epsilon=st.floats(
             min_value=0.01, max_value=0.99, allow_nan=False, allow_infinity=False
         ),
-        data_strategy=st.data(),
         **hu.gcs
     )
-    def test_row_wise_sparse_adagrad(self, inputs, lr, epsilon, data_strategy, gc, dc):
+    def test_row_wise_sparse_adagrad(self, inputs, lr, epsilon, gc, dc):
         adagrad_sparse_test_helper(
             self,
             inputs,
@@ -200,11 +198,10 @@ def test_row_wise_sparse_adagrad(self, inputs, lr, epsilon, data_strategy, gc, d
         epsilon=st.floats(
             min_value=0.01, max_value=0.99, allow_nan=False, allow_infinity=False
         ),
-        data_strategy=st.data(),
         **hu.gcs
     )
     def test_row_wise_sparse_adagrad_empty(
-        self, inputs, lr, epsilon, data_strategy, gc, dc
+        self, inputs, lr, epsilon, gc, dc
     ):
         param, momentum = inputs
         grad = np.empty(shape=(0,) + param.shape[1:], dtype=np.float32)

caffe2/python/operator_test/wngrad_test.py

@@ -182,10 +182,8 @@ def test_sparse_wngrad(self, inputs, seq_b, lr, epsilon, gc, dc):
                         allow_nan=False, allow_infinity=False),
            epsilon=st.floats(min_value=0.01, max_value=0.99,
                              allow_nan=False, allow_infinity=False),
-           data_strategy=st.data(),
            **hu.gcs_cpu_only)
-    def test_sparse_wngrad_empty(self, inputs, seq_b, lr, epsilon,
-                                  data_strategy, gc, dc):
+    def test_sparse_wngrad_empty(self, inputs, seq_b, lr, epsilon, gc, dc):
         param = inputs[0]
         seq_b = np.array([seq_b, ], dtype=np.float32)
         lr = np.array([lr], dtype=np.float32)