[CodeStyle][E262][E265] make comments start with `# ` (#47687)

* [CodeStyle][E262][E265] make comments start with `# `

* flake8 config

.flake8

@@ -16,7 +16,7 @@ exclude =
     ./python/paddle/fluid/tests/unittests/mlu/**
 ignore =
     # E, see https://pycodestyle.pycqa.org/en/latest/intro.html#error-codes
-    E203,E262,E265,E266,
+    E203,E266,
     E401,E402,
     E501,
     E721,E722,E731,E741,

python/paddle/distributed/fleet/meta_optimizers/ascend/ascend_parser.py

@@ -17,7 +17,7 @@ from functools import reduce
 
 __all__ = []
 
-registerd_op = {  ## forwards
+registerd_op = {  # forwards
     "elementwise_add": "AddParser",
     "matmul": "MatMulParser",
     "mul": "MulParser",

python/paddle/distributed/fleet/meta_optimizers/parameter_server_optimizer.py

@@ -394,7 +394,7 @@ class ParameterServerOptimizer(MetaOptimizerBase):
                 loss.block.program._heter_pipeline_opt = {
                     "trainer": "HeterPipelineTrainer",
                     "device_worker": "HeterSection",
-                    "trainers": self.role_maker._get_stage_trainers(),  ## trainer num in each stage
+                    "trainers": self.role_maker._get_stage_trainers(),  # trainer num in each stage
                     "trainer_id": int(self.role_maker._role_id()),
                     "pipeline_stage": int(self.role_maker._get_stage_id()) - 1,
                     "num_pipeline_stages": int(

python/paddle/distributed/passes/auto_parallel_grad_clip.py

@@ -269,7 +269,7 @@ class ClipGradByGloblNormPass(PassBase):
             if op.type in removed_op_out_type:
                 input_name = op.input("X")[0]
                 if input_name.find("@GRAD") != -1:
-                    #'clip_by_norm', 'squared_l2_norm', 'square'
+                    # 'clip_by_norm', 'squared_l2_norm', 'square'
                     param_name = input_name[: input_name.find("@GRAD")]
                     is_local = self.clip_helper._is_local_param(param_name)
                     is_calculate = self.clip_helper._is_calcuate_norm(

python/paddle/distributed/passes/ps_trainer_pass.py

@@ -1222,8 +1222,8 @@ class SplitTrainerOpsPass(PassBase):
             # runtime attribute
             "endpoint": get_trainer_endpoint(
                 role_maker
-            ),  ## get trainer endpoint
-            "fanin": 0,  ## get heter worker
+            ),  # get trainer endpoint
+            "fanin": 0,  # get heter worker
             "pserver_id": get_role_id(role_maker),
             "distributed_mode": attrs['ps_mode'],
             "rpc_exec_thread_num": int(os.getenv("CPU_NUM", 32)),
@@ -1296,7 +1296,7 @@ class SetHeterPipelineOptPass(PassBase):
         main_program._heter_pipeline_opt = {
             "trainer": "HeterPipelineTrainer",
             "device_worker": "HeterSection",
-            "trainers": role_maker._get_stage_trainers(),  ## trainer num in each stage
+            "trainers": role_maker._get_stage_trainers(),  # trainer num in each stage
             "trainer_id": int(role_maker._role_id()),
             "pipeline_stage": int(role_maker._get_stage_id()) - 1,
             "num_pipeline_stages": int(role_maker._get_num_stage()),
@@ -1524,7 +1524,7 @@ class SplitFlOpsPass(PassBase):
         attrs = {
             "message_to_block_id": [grad_to_block_id],
             "optimize_blocks": [second_block],
-            "endpoint": get_trainer_endpoint(self.role_maker),  ##
+            "endpoint": get_trainer_endpoint(self.role_maker),
             "fanin": 0,
             "pserver_id": get_role_id(self.role_maker),
             "distributed_mode": self.ps_mode,
@@ -1584,7 +1584,7 @@ class SplitFlOpsPass(PassBase):
         grad_to_block_id = block_input_flag + ":" + str(second_block.idx)
         attrs = {
             "message_to_block_id": [grad_to_block_id],
-            "optimize_blocks": [second_block],  ## what to do?
+            "optimize_blocks": [second_block],  # what to do?
             "endpoint": get_heter_worker_endpoint(self.role_maker),
             "fanin": len(get_previous_stage_trainers(self.role_maker)),
             "pserver_id": 1,  # TODO

python/paddle/fluid/tests/unittests/dist_fleet_ctr.py

@@ -331,7 +331,7 @@ class TestDistCTR2x2(FleetDistRunnerBase):
         dataset.set_pipe_command('python ctr_dataset_reader.py')
         dataset.load_into_memory()
 
-        dataset.global_shuffle(fleet, 12)  ##TODO: thread configure
+        dataset.global_shuffle(fleet, 12)  # TODO: thread configure
         shuffle_data_size = dataset.get_shuffle_data_size(fleet)
         local_data_size = dataset.get_shuffle_data_size()
         data_size_list = fleet.util.all_gather(local_data_size)

python/paddle/fluid/tests/unittests/ps/ps_dnn_trainer.py

@@ -172,7 +172,7 @@ def get_user_defined_strategy(config):
             micro_num = 1
             strategy.pipeline_configs = {
                 "accumulate_steps": micro_num
-            }  ## num_microbatches
+            }  # num_microbatches
     elif sync_mode == "geo":
         strategy = paddle.distributed.fleet.DistributedStrategy()
         strategy.a_sync = True
@@ -372,7 +372,7 @@ class DnnTrainer(object):
             print("entering run_minimize -- old")
             fleet_obj = fleet.distributed_optimizer(
                 inner_optimizer, user_defined_strategy
-            )  ## Fleet 对象
+            )  # Fleet object
             fleet_obj.minimize(loss)
 
         if fleet.is_server():

python/paddle/fluid/tests/unittests/test_cross_entropy_loss.py

@@ -36,8 +36,8 @@ def cross_entropy_loss_1d(
     C = input_shape[1]
     out = np.zeros_like(label).astype(np.float64)
     total_weight = 0
-    ###1. compute softmax cross_entropy (with weight)
-    ###   Note: only support hard labels.
+    # 1. compute softmax cross_entropy (with weight)
+    #    Note: only support hard labels.
     for i in range(N):
         cur_target = label[i]
         if cur_target == ignore_index:
@@ -47,7 +47,7 @@ def cross_entropy_loss_1d(
         total_weight += cur_weight
         out[i] = -log_softmax_out[i][cur_target] * cur_weight
 
-    ###2. deal with reduction
+    # 2. deal with reduction
     if reduction == 'sum':
         return np.sum(out), np.array([total_weight]).astype('float64')
     elif reduction == 'mean':
@@ -179,7 +179,7 @@ class CrossEntropyLoss(unittest.TestCase):
             'float32' if fluid.core.is_compiled_with_rocm() else 'float64'
         )
 
-    ###test for deprecated softmax_with_cross_entropy
+    # test for deprecated softmax_with_cross_entropy
     def test_softmax_with_cross_entropy(self):
         self.numeric_stable_mode = False
         self.soft_label = True
@@ -240,8 +240,8 @@ class CrossEntropyLoss(unittest.TestCase):
         )
         np.testing.assert_allclose(paddle_loss_ce.numpy(), expected, rtol=1e-05)
 
-    ###soft_label test start
-    ###soft_label test 1
+    # soft_label test start
+    # soft_label test 1
     def test_cross_entropy_loss_soft_1d(self):
         self.numeric_stable_mode = False
         self.soft_label = True
@@ -329,7 +329,7 @@ class CrossEntropyLoss(unittest.TestCase):
         np.testing.assert_allclose(static_ret[0], expected, rtol=1e-05)
         np.testing.assert_allclose(dy_ret_value, expected, rtol=1e-05)
 
-    ###soft_label test 2
+    # soft_label test 2
     def test_cross_entropy_loss_soft_1d_weight(self):
         self.numeric_stable_mode = False
         self.soft_label = True
@@ -427,7 +427,7 @@ class CrossEntropyLoss(unittest.TestCase):
         np.testing.assert_allclose(static_ret[0], expected, rtol=1e-05)
         np.testing.assert_allclose(dy_ret_value, expected, rtol=1e-05)
 
-    ###soft_label test 3
+    # soft_label test 3
     def test_cross_entropy_loss_soft_1d_mean(self):
         self.numeric_stable_mode = False
         self.soft_label = True
@@ -511,7 +511,7 @@ class CrossEntropyLoss(unittest.TestCase):
         np.testing.assert_allclose(static_ret[0], expected, rtol=1e-05)
         np.testing.assert_allclose(dy_ret_value, expected, rtol=1e-05)
 
-    ###soft_label test 4
+    # soft_label test 4
     def test_cross_entropy_loss_soft_1d_weight_mean(self):
         self.numeric_stable_mode = False
         self.soft_label = True
@@ -599,7 +599,7 @@ class CrossEntropyLoss(unittest.TestCase):
         np.testing.assert_allclose(static_ret[0], expected, rtol=1e-05)
         np.testing.assert_allclose(dy_ret_value, expected, rtol=1e-05)
 
-    ###soft_label test 5
+    # soft_label test 5
     def test_cross_entropy_loss_soft_2d(self):
         def inner_cross_entropy_loss_soft_2d(soft_label):
             self.numeric_stable_mode = False
@@ -704,7 +704,7 @@ class CrossEntropyLoss(unittest.TestCase):
         inner_cross_entropy_loss_soft_2d(True)
         inner_cross_entropy_loss_soft_2d(False)
 
-    ###soft_label test 6
+    # soft_label test 6
     def test_cross_entropy_loss_soft_2d_weight_mean(self):
         self.numeric_stable_mode = False
         self.soft_label = True
@@ -801,7 +801,7 @@ class CrossEntropyLoss(unittest.TestCase):
         np.testing.assert_allclose(static_ret[0], expected, rtol=1e-05)
         np.testing.assert_allclose(dy_ret_value, expected, rtol=1e-05)
 
-    ###soft_label test end
+    # soft_label test end
 
     def test_cross_entropy_loss_1d_with_mean_ignore(self):
         input_np = np.random.random([2, 4]).astype(self.dtype)

python/paddle/fluid/tests/unittests/test_fused_feedforward_op.py

@@ -253,7 +253,7 @@ class APITestStaticFusedFFN(unittest.TestCase):
             pre_layer_norm=False,
         )
 
-        ######base ffn######
+        # base ffn
         linear1_out = F.linear(x, linear1_weight, linear1_bias)
         act_out = F.relu(linear1_out)
         dropout1_out = F.dropout(x=act_out, p=0.0, training=False)
@@ -265,7 +265,6 @@ class APITestStaticFusedFFN(unittest.TestCase):
             weight=ln2_scale,
             bias=ln2_bias,
         )
-        ######base ffn######
 
         exe = paddle.static.Executor(paddle.CUDAPlace(0))
 

python/paddle/fluid/tests/unittests/test_get_inputs_outputs_in_block.py

@@ -68,9 +68,9 @@ class TestGetInputsOutputsInBlock(unittest.TestCase):
         inner_inputs, inner_outputs = utils.get_inputs_outputs_in_block(
             sub_block
         )
-        #'fill_constant_1.tmp_0', 'tmp_3' are names of a, c
+        # 'fill_constant_1.tmp_0', 'tmp_3' are names of a, c
         self.assertTrue(inner_inputs == {'fill_constant_1.tmp_0', 'tmp_3'})
-        #'_generated_var_1', is name of a + c
+        # '_generated_var_1', is name of a + c
         self.assertTrue(inner_outputs == {'_generated_var_1'})
 
 

python/paddle/fluid/tests/unittests/test_multi_dot_op.py

@@ -209,7 +209,7 @@ class TestMultiDotOp4MatFirstAndLast1D(TestMultiDotOp4Mat):
         self.outputs = {'Out': multi_dot([self.A, self.B, self.C, self.D])}
 
 
-#####python API test#######
+# python API test
 class TestMultiDotOpError(unittest.TestCase):
     def test_errors(self):
         with paddle.static.program_guard(

python/paddle/fluid/tests/unittests/test_slice_op.py

@@ -298,7 +298,7 @@ class TestSliceOp_decs_dim_starts_OneTensor(OpTest):
         self.outputs = {'Out': self.out}
         self.attrs = {
             'axes': self.axes,
-            #'starts': self.starts,
+            # 'starts': self.starts,
             'ends': self.ends,
             'infer_flags': self.infer_flags,
             'decrease_axis': self.decrease_axis,
@@ -335,8 +335,8 @@ class TestSliceOp_starts_OneTensor_ends_OneTensor(OpTest):
         self.outputs = {'Out': self.out}
         self.attrs = {
             'axes': self.axes,
-            #'starts': self.starts,
-            #'ends': self.ends_infer,
+            # 'starts': self.starts,
+            # 'ends': self.ends_infer,
             'infer_flags': self.infer_flags,
         }
 
@@ -369,8 +369,8 @@ class TestSliceOp_decs_dim_starts_and_ends_OneTensor(OpTest):
         self.outputs = {'Out': self.out}
         self.attrs = {
             'axes': self.axes,
-            #'starts': self.starts,
-            #'ends': self.ends,
+            # 'starts': self.starts,
+            # 'ends': self.ends,
             'infer_flags': self.infer_flags,
             'decrease_axis': self.decrease_axis,
         }
@@ -412,7 +412,7 @@ class TestSliceOp_starts_OneTensor_ends_ListTensor(OpTest):
         self.outputs = {'Out': self.out}
         self.attrs = {
             'axes': self.axes,
-            #'starts': self.starts,
+            # 'starts': self.starts,
             'ends': self.ends_infer,
             'infer_flags': self.infer_flags,
         }

python/paddle/fluid/tests/unittests/test_strided_slice_op.py

@@ -406,7 +406,7 @@ class TestStridedSliceOp_starts_Tensor(OpTest):
         self.outputs = {'Out': self.output}
         self.attrs = {
             'axes': self.axes,
-            #'starts': self.starts,
+            # 'starts': self.starts,
             'ends': self.ends,
             'strides': self.strides,
             'infer_flags': self.infer_flags,
@@ -442,7 +442,7 @@ class TestStridedSliceOp_ends_Tensor(OpTest):
         self.attrs = {
             'axes': self.axes,
             'starts': self.starts,
-            #'ends': self.ends,
+            # 'ends': self.ends,
             'strides': self.strides,
             'infer_flags': self.infer_flags,
         }
@@ -483,8 +483,8 @@ class TestStridedSliceOp_listTensor_Tensor(OpTest):
         self.outputs = {'Out': self.output}
         self.attrs = {
             'axes': self.axes,
-            #'starts': self.starts,
-            #'ends': self.ends,
+            # 'starts': self.starts,
+            # 'ends': self.ends,
             'strides': self.strides,
             'infer_flags': self.infer_flags,
         }
@@ -520,7 +520,7 @@ class TestStridedSliceOp_strides_Tensor(OpTest):
             'axes': self.axes,
             'starts': self.starts,
             'ends': self.ends,
-            #'strides': self.strides,
+            # 'strides': self.strides,
             'infer_flags': self.infer_flags,
         }
 

python/paddle/fluid/tests/unittests/white_list/check_op_sequence_instance_0_input_white_list.py

@@ -18,7 +18,6 @@
 # compiletime&runtime will be skipped. Ops in this whitelist need to declear
 # reasons for skipping compile_vs_runtime test or be fixed later.
 
-#!/usr/bin/env python
 import sys
 
 # For ops in this whitelist, the check of instance size is 0 input will be skipped.

python/paddle/framework/__init__.py

@@ -47,9 +47,9 @@ from ..fluid.framework import set_flags  # noqa: F401
 from ..fluid.dygraph.base import enable_dygraph as disable_static  # noqa: F401
 from ..fluid.dygraph.base import disable_dygraph as enable_static  # noqa: F401
 from ..fluid.framework import _non_static_mode as in_dynamic_mode  # noqa: F401
-from ..fluid.framework import (
-    _non_static_mode,
-)  #  noqa: F401; temporary used for hackson
+from ..fluid.framework import (  # noqa: F401
+    _non_static_mode,  # temporary used for hackson
+)
 from ..fluid.framework import (
     _current_expected_place,
     _get_paddle_place,

tools/parallel_UT_rule.py

@@ -228,7 +228,7 @@ HIGH_PARALLEL_JOB_NEW = [
     'test_launch_coverage',
     'test_mkldnn_conv_activation_fuse_pass',
     'test_inference_model_io',
-    'test_fusion_repeated_fc_relu_op',  #'heter_listen_and_server_test',
+    'test_fusion_repeated_fc_relu_op',
     'cudnn_desc_test',
     'test_beam_search_op',
     'test_var_conv_2d',

tools/sampcd_processor.py

@@ -547,7 +547,7 @@ def get_full_api():
     """
     get all the apis
     """
-    global API_DIFF_SPEC_FN  ## readonly
+    global API_DIFF_SPEC_FN  # readonly
     from print_signatures import get_all_api_from_modulelist
 
     member_dict = get_all_api_from_modulelist()
@@ -559,7 +559,7 @@ def get_full_api_by_walk():
     """
     get all the apis
     """
-    global API_DIFF_SPEC_FN  ## readonly
+    global API_DIFF_SPEC_FN  # readonly
     from print_signatures import get_all_api
 
     apilist = get_all_api()
@@ -571,7 +571,7 @@ def get_full_api_from_pr_spec():
     """
     get all the apis
     """
-    global API_PR_SPEC_FN, API_DIFF_SPEC_FN  ## readonly
+    global API_PR_SPEC_FN, API_DIFF_SPEC_FN  # readonly
     pr_api = get_api_md5(API_PR_SPEC_FN)
     if len(pr_api):
         with open(API_DIFF_SPEC_FN, 'w') as f:
@@ -584,7 +584,7 @@ def get_incrementapi():
     '''
     this function will get the apis that difference between API_DEV.spec and API_PR.spec.
     '''
-    global API_DEV_SPEC_FN, API_PR_SPEC_FN, API_DIFF_SPEC_FN  ## readonly
+    global API_DEV_SPEC_FN, API_PR_SPEC_FN, API_DIFF_SPEC_FN  # readonly
     dev_api = get_api_md5(API_DEV_SPEC_FN)
     pr_api = get_api_md5(API_PR_SPEC_FN)
     with open(API_DIFF_SPEC_FN, 'w') as f: