Fix some typos (dtyte, paramerter, etc.)

test/legacy_test/test_adadelta_op.py

@@ -236,7 +236,7 @@ def test_adadelta_dygraph(self):
         adam.clear_gradients()
 
 
-class TestAdadeltaOpMultiPrecison(unittest.TestCase):
+class TestAdadeltaOpMultiPrecision(unittest.TestCase):
     def _test_adadelta_op_dygraph_place_amp(self, place, use_amp=False):
         import paddle
 

test/legacy_test/test_adagrad_op.py

@@ -130,7 +130,7 @@ def check_with_place(self, place):
         param_array = np.full((height, row_numel), 5.0).astype("float32")
         param.set(param_array, place)
 
-        # create and initialize LeraningRate Variable
+        # create and initialize LearningRate Variable
         lr = scope.var('LearningRate').get_tensor()
         lr_array = np.full((1), 2.0).astype("float32")
         lr.set(lr_array, place)
@@ -211,7 +211,7 @@ def test_sparse_adagrad(self):
             self.check_with_place(place)
 
 
-class TestAdagradOpMultiPrecison(unittest.TestCase):
+class TestAdagradOpMultiPrecision(unittest.TestCase):
     def _test_adagrad_op_dygraph_place_amp(self, place, use_amp=False):
         import paddle
 

test/legacy_test/test_adam_op.py

@@ -758,7 +758,7 @@ def test_adam_op_with_state_dict(self):
         state_dict = adam.state_dict()
         adam.set_state_dict(state_dict)
 
-        # leanrning_rate is Tensor
+        # learning_rate is Tensor
         with self.assertRaises(TypeError):
             learning_rate = np.array([0.01]).astype("float32")
             learning_rate = paddle.to_tensor(learning_rate)

test/legacy_test/test_adamax_op.py

@@ -239,7 +239,7 @@ def test_adamax_op_invalid_input(self):
             )
 
 
-class TestAdamaxOpMultiPrecison(unittest.TestCase):
+class TestAdamaxOpMultiPrecision(unittest.TestCase):
     def _test_adamax_op_dygraph_place_amp(self, place, use_amp=False):
         import paddle
 

test/legacy_test/test_adamw_op.py

@@ -404,7 +404,7 @@ def test_adamw_op_dygraph_bypassing_step(self):
             adam.clear_gradients()
 
 
-class TestAdamWOpMultiPrecisonWithMainGrad(unittest.TestCase):
+class TestAdamWOpMultiPrecisionWithMainGrad(unittest.TestCase):
     def _test_adamw_op_dygraph_place_amp_with_maingrad(
         self, place, shape, use_main_grad
     ):
@@ -543,7 +543,7 @@ def test_main(self):
                     )
 
 
-class TestAdamWOpMultiPrecison(unittest.TestCase):
+class TestAdamWOpMultiPrecision(unittest.TestCase):
     def _test_adamw_op_dygraph_place_amp(self, place, use_amp=False):
         paddle.disable_static()
         paddle.seed(10)

test/legacy_test/test_add_n_op.py

@@ -70,15 +70,15 @@ def test_add_n_api(self):
         if not paddle.is_compiled_with_cuda():
             return
         dtypes = ['float32', 'complex64', 'complex128']
-        for dtyte in dtypes:
-            if dtyte == 'complex64' or dtyte == 'complex128':
+        for dtype in dtypes:
+            if dtype == 'complex64' or dtype == 'complex128':
                 self.x_np = (
                     np.random.random([self.l, 16, 256])
                     + 1j * np.random.random([self.l, 16, 256])
-                ).astype(dtyte)
+                ).astype(dtype)
 
-            y_np_32, x_g_np_32 = self.check_main(self.x_np, dtyte)
-            y_np_gt = np.sum(self.x_np, axis=0).astype(dtyte)
+            y_np_32, x_g_np_32 = self.check_main(self.x_np, dtype)
+            y_np_gt = np.sum(self.x_np, axis=0).astype(dtype)
             np.testing.assert_allclose(y_np_32, y_np_gt, rtol=1e-06)
 
 

test/legacy_test/test_arange.py

@@ -59,7 +59,7 @@ def init_config(self):
         self.case = (0, 5, 1)
 
 
-class TestFloa16ArangeOp(TestArangeOp):
+class TestFloat16ArangeOp(TestArangeOp):
     def init_config(self):
         self.dtype = np.float16
         self.python_api = paddle.arange

test/legacy_test/test_backward.py

@@ -378,12 +378,12 @@ def callback(block, context):
 
 
 class TestGradientsWithOptimizer(unittest.TestCase):
-    def _check_grad_op_name(self, forward_list, optimiezed_list):
+    def _check_grad_op_name(self, forward_list, optimized_list):
         backward_list = [op + "_grad" for op in reversed(forward_list)]
-        idx = optimiezed_list.index(backward_list[0], len(backward_list))
+        idx = optimized_list.index(backward_list[0], len(backward_list))
 
         self.assertListEqual(
-            backward_list, optimiezed_list[idx : idx + len(backward_list)]
+            backward_list, optimized_list[idx : idx + len(backward_list)]
         )
 
     def test_gradient_with_optimizer(self):

test/legacy_test/test_bicubic_interp_v2_op.py

@@ -674,7 +674,7 @@ def test_case(self):
 
 class TestBicubicOpError(unittest.TestCase):
     def test_imperative_errors(self):
-        # the input of interpoalte must be Variable.
+        # the input of interpolate must be Variable.
         x1 = base.create_lod_tensor(
             np.array([-1, 3, 5, 5]), [[1, 1, 1, 1]], base.CPUPlace()
         )
@@ -687,7 +687,7 @@ def test_mode_type():
             )
 
             out = interpolate(
-                x, size=[12, 12], mode='UNKONWN', align_corners=False
+                x, size=[12, 12], mode='UNKNOWN', align_corners=False
             )
 
         def test_input_shape():
@@ -696,7 +696,7 @@ def test_input_shape():
                 x, size=[12, 12], mode='BICUBIC', align_corners=False
             )
 
-        def test_align_corcers():
+        def test_align_corners():
             x = paddle.static.data(
                 name="x", shape=[2, 3, 6, 6], dtype="float32"
             )
@@ -887,7 +887,7 @@ def test_input_shape_1():
 
         self.assertRaises(ValueError, test_mode_type)
         self.assertRaises(ValueError, test_input_shape)
-        self.assertRaises(TypeError, test_align_corcers)
+        self.assertRaises(TypeError, test_align_corners)
         self.assertRaises(ValueError, test_attr_data_format)
         self.assertRaises(TypeError, test_actual_shape)
         self.assertRaises(ValueError, test_scale_value)

test/legacy_test/test_checkpoint_saver.py

@@ -18,7 +18,7 @@
 from paddle.distributed.fleet.utils.fs import HDFSClient
 
 
-class CheckpointerSaverTest(unittest.TestCase):
+class CheckpointSaverTest(unittest.TestCase):
     def test(self):
         fs = HDFSClient("/usr/local/hadoop-2.7.7", None)
         dir_path = "./checkpointsaver_test"

test/legacy_test/test_collective_api_base.py

@@ -611,19 +611,19 @@ def convertbf16(origin):
             result1 = []
             result2 = []
 
-            def is_empyt_list(x):
+            def is_empty_list(x):
                 if isinstance(x, list) and len(x) == 0:
                     return True
                 return False
 
             for i in range(tot_expert):
                 for arr in output1[i]:
-                    if is_empyt_list(arr):
+                    if is_empty_list(arr):
                         continue
                     result1.append(arr)
             for i in range(tot_expert):
                 for arr in output2[i]:
-                    if is_empyt_list(arr):
+                    if is_empty_list(arr):
                         continue
                     result2.append(arr)
 

test/legacy_test/test_conv2d_api.py

@@ -201,7 +201,7 @@ def run_5():
 
         self.assertRaises(ValueError, run_5)
 
-        # ValueError: channel dimmention
+        # ValueError: channel dimension
         x = paddle.static.data(
             name="x",
             shape=[2, 5, 5, -1],

test/legacy_test/test_conv3d_layer.py

@@ -27,7 +27,7 @@ def __init__(
         self,
         methodName='runTest',
         batch_size=4,
-        spartial_shape=(8, 8, 8),
+        spatial_shape=(8, 8, 8),
         num_channels=6,
         num_filters=8,
         filter_size=3,
@@ -43,7 +43,7 @@ def __init__(
         self.batch_size = batch_size
         self.num_channels = num_channels
         self.num_filters = num_filters
-        self.spartial_shape = spartial_shape
+        self.spatial_shape = spatial_shape
         self.filter_size = filter_size
 
         self.padding = padding
@@ -58,13 +58,13 @@ def setUp(self):
         self.channel_last = self.data_format == "NDHWC"
         if self.channel_last:
             input_shape = (
-                (self.batch_size,) + self.spartial_shape + (self.num_channels,)
+                (self.batch_size,) + self.spatial_shape + (self.num_channels,)
             )
         else:
             input_shape = (
                 self.batch_size,
                 self.num_channels,
-            ) + self.spartial_shape
+            ) + self.spatial_shape
         self.input = np.random.randn(*input_shape).astype(self.dtype)
 
         if isinstance(self.filter_size, int):

test/legacy_test/test_conv3d_op.py

@@ -1212,7 +1212,7 @@ def run_5():
 
             self.assertRaises(ValueError, run_5)
 
-            # ValueError: channel dimmention
+            # ValueError: channel dimension
             x = paddle.static.data(
                 name="x",
                 shape=[2, 5, 5, 5, -1],

test/legacy_test/test_conv3d_transpose_layer.py

@@ -27,7 +27,7 @@ def __init__(
         self,
         methodName='runTest',
         batch_size=2,
-        spartial_shape=(8, 8, 8),
+        spatial_shape=(8, 8, 8),
         num_channels=6,
         num_filters=8,
         filter_size=3,
@@ -44,7 +44,7 @@ def __init__(
         self.batch_size = batch_size
         self.num_channels = num_channels
         self.num_filters = num_filters
-        self.spartial_shape = spartial_shape
+        self.spatial_shape = spatial_shape
         self.filter_size = filter_size
         self.output_size = output_size
 
@@ -60,13 +60,13 @@ def setUp(self):
         self.channel_last = self.data_format == "NDHWC"
         if self.channel_last:
             input_shape = (
-                (self.batch_size,) + self.spartial_shape + (self.num_channels,)
+                (self.batch_size,) + self.spatial_shape + (self.num_channels,)
             )
         else:
             input_shape = (
                 self.batch_size,
                 self.num_channels,
-            ) + self.spartial_shape
+            ) + self.spatial_shape
         self.input = np.random.randn(*input_shape).astype(self.dtype)
 
         if isinstance(self.filter_size, int):

test/legacy_test/test_fused_attention_op_api.py

@@ -53,9 +53,9 @@ def layer_norm(x, has_scale, has_bias, weight, bias, epsilon=1e-05):
     batch_size, src_len, d_model = x.shape
     x = x.reshape((batch_size * src_len, d_model))
     mu = np.mean(x, axis=1, keepdims=True)
-    sigma_squar = np.sum(np.square(x - mu), axis=1) / d_model
+    sigma_square = np.sum(np.square(x - mu), axis=1) / d_model
     x1_up = x - mu
-    x1_down_1 = sigma_squar + epsilon
+    x1_down_1 = sigma_square + epsilon
     x1_down = np.sqrt(x1_down_1)
     x1_down = x1_down.reshape((x1_down.shape[0], 1))
     x1 = x1_up / x1_down

test/legacy_test/test_fused_bias_dropout_residual_layer_norm_op_api.py

@@ -27,9 +27,9 @@ def layer_norm(x, has_scale, has_bias, weight, bias, epsilon=1e-05):
     batch_size, src_len, d_model = x.shape
     x = x.reshape((batch_size * src_len, d_model))
     mu = np.mean(x, axis=1, keepdims=True)
-    sigma_squar = np.sum(np.square(x - mu), axis=1) / d_model
+    sigma_square = np.sum(np.square(x - mu), axis=1) / d_model
     x1_up = x - mu
-    x1_down_1 = sigma_squar + epsilon
+    x1_down_1 = sigma_square + epsilon
     x1_down = np.sqrt(x1_down_1)
     x1_down = x1_down.reshape((x1_down.shape[0], 1))
     x1 = x1_up / x1_down

test/legacy_test/test_log_softmax.py

@@ -194,7 +194,7 @@ def check_api(self, axis=-1):
             out = exe.run(feed={'x': self.x}, fetch_list=[y])
         np.testing.assert_allclose(out[0], ref_out, rtol=1e-05)
 
-        # test dygrapg api
+        # test dygraph api
         paddle.disable_static()
         x = paddle.to_tensor(self.x)
         y = logsoftmax(x)

test/legacy_test/test_paddle_save_load_binary.py

@@ -94,7 +94,7 @@ def test_replace_save_load_vars(self):
                     t = np.array(
                         base.global_scope().find_var(var.name).get_tensor()
                     )
-                    # make sure all the paramerter or optimizer var have been update
+                    # make sure all the parameter or optimizer var have been update
                     self.assertTrue(np.sum(np.abs(t)) != 0)
                     base_map[var.name] = t
             # test for replace_save_vars/io.load_vars

test/legacy_test/test_scatter_nd_op.py

@@ -29,16 +29,16 @@ def numpy_scatter_nd(ref, index, updates, fun):
     index_shape = index.shape
 
     end_size = index_shape[-1]
-    remain_numl = 1
+    remain_numel = 1
     for i in range(len(index_shape) - 1):
-        remain_numl *= index_shape[i]
+        remain_numel *= index_shape[i]
 
     slice_size = 1
     for i in range(end_size, len(ref_shape)):
         slice_size *= ref_shape[i]
 
-    flat_index = index.reshape([remain_numl] + list(index_shape[-1:]))
-    flat_updates = updates.reshape((remain_numl, slice_size))
+    flat_index = index.reshape([remain_numel] + list(index_shape[-1:]))
+    flat_updates = updates.reshape((remain_numel, slice_size))
     flat_output = ref.reshape(list(ref_shape[:end_size]) + [slice_size])
 
     for i_up, i_out in enumerate(flat_index):