self.cdata = storage._weak_ref()
        # Save a direct reference to _free_weak_ref because the `torch` module
        # might be cleared during Python shutdown before this module is cleared.
        self._free_weak_ref = torch.Storage._free_weak_ref  # type: ignore[attr-defined]

    def expired(self):
        return torch.Storage._expired(self.cdata)  # type: ignore[attr-defined]

    def __del__(self):

            

Reported by Pylint.

                      self._free_weak_ref = torch.Storage._free_weak_ref  # type: ignore[attr-defined]

    def expired(self):
        return torch.Storage._expired(self.cdata)  # type: ignore[attr-defined]

    def __del__(self):
        self._free_weak_ref(self.cdata)



            

Reported by Pylint.

              

def reduce_storage(storage):
    from . import get_sharing_strategy
    if storage.is_cuda:
        raise RuntimeError("Cannot pickle CUDA storage; try pickling a CUDA tensor instead")
    elif get_sharing_strategy() == 'file_system':
        metadata = storage._share_filename_()
        cache_key = metadata[1]

            

Reported by Pylint.

                      self._free_weak_ref = torch.Storage._free_weak_ref  # type: ignore[attr-defined]

    def expired(self):
        return torch.Storage._expired(self.cdata)  # type: ignore[attr-defined]

    def __del__(self):
        self._free_weak_ref(self.cdata)



            

Reported by Pylint.

              class SharedCache(dict):
    """dictionary from multiprocessing handles to StorageWeakRef"""

    def __init__(self):
        # free_dead_references() is called if the len exceeds the current
        # limit. The limit scales with the number of remaining live objects.
        self.limit = 128
        # `fork` inherits lock state, so in case we fork when the lock is held,
        # we register a function to reset the lock to a new object to avoid

            

Reported by Pylint.

              
def rebuild_tensor(cls, storage, metadata):
    storage_offset, size, stride, requires_grad = metadata
    t = torch._utils._rebuild_tensor(storage, storage_offset, size, stride)
    if cls == torch.nn.parameter.Parameter:
        # we have to pass requires_grad into constructor, rather than set it as an
        # attribute later, because it's an important check for Integer Tensors to
        # have requires_grad=False (or else they raise an error)
        t = torch.nn.parameter.Parameter(t, requires_grad=requires_grad)

            

Reported by Pylint.

              
def rebuild_tensor(cls, storage, metadata):
    storage_offset, size, stride, requires_grad = metadata
    t = torch._utils._rebuild_tensor(storage, storage_offset, size, stride)
    if cls == torch.nn.parameter.Parameter:
        # we have to pass requires_grad into constructor, rather than set it as an
        # attribute later, because it's an important check for Integer Tensors to
        # have requires_grad=False (or else they raise an error)
        t = torch.nn.parameter.Parameter(t, requires_grad=requires_grad)

            

Reported by Pylint.

                  else:
        storage = storage_from_cache(storage_cls, (storage_handle, storage_offset_bytes))
        if storage is None:
            torch.cuda._lazy_init()
            storage = storage_cls._new_shared_cuda(
                storage_device,
                storage_handle,
                storage_size_bytes,
                storage_offset_bytes,

            

Reported by Pylint.

                      storage = storage_from_cache(storage_cls, (storage_handle, storage_offset_bytes))
        if storage is None:
            torch.cuda._lazy_init()
            storage = storage_cls._new_shared_cuda(
                storage_device,
                storage_handle,
                storage_size_bytes,
                storage_offset_bytes,
                ref_counter_handle,

            

Reported by Pylint.

                          shared_cache[(storage_handle, storage_offset_bytes)] = StorageWeakRef(storage)
        else:
            # We already ref counting this Storage, but producer needs new ref-counters to be released.
            storage_cls._release_ipc_counter(ref_counter_handle, ref_counter_offset)

    t = torch._utils._rebuild_tensor(storage, tensor_offset, tensor_size, tensor_stride)

    if tensor_cls == torch.nn.parameter.Parameter:
        # It is crucial for integer tensors to receive

            

Reported by Pylint.

              from caffe2.proto import caffe2_pb2

from functools import partial
from hypothesis import given
from hypothesis import settings as ht_settings
import hypothesis.strategies as st
import numpy as np
import unittest


            

Reported by Pylint.

              
from functools import partial
from hypothesis import given
from hypothesis import settings as ht_settings
import hypothesis.strategies as st
import numpy as np
import unittest



            

Reported by Pylint.

              from functools import partial
from hypothesis import given
from hypothesis import settings as ht_settings
import hypothesis.strategies as st
import numpy as np
import unittest


def gru_unit(*args, **kwargs):

            

Reported by Pylint.

                  return (hidden_t, )


def gru_reference(input, hidden_input,
                  reset_gate_w, reset_gate_b,
                  update_gate_w, update_gate_b,
                  output_gate_w, output_gate_b,
                  seq_lengths, drop_states=False,
                  linear_before_reset=False):

            

Reported by Pylint.

                  return dims_.flatmap(create_input)


def _prepare_gru_unit_op(gc, n, d, outputs_with_grads,
                         forward_only=False, drop_states=False,
                         sequence_lengths=False,
                         two_d_initial_states=None):
    print("Dims: (n,d) = ({},{})".format(n, d))


            

Reported by Pylint.

              

def _prepare_gru_unit_op(gc, n, d, outputs_with_grads,
                         forward_only=False, drop_states=False,
                         sequence_lengths=False,
                         two_d_initial_states=None):
    print("Dims: (n,d) = ({},{})".format(n, d))

    def generate_input_state(n, d):

            

Reported by Pylint.

                      **hu.gcs
    )
    def test_gru_unit_op(self, seed, input_tensor, fwd_only,
                         drop_states, sequence_lengths, gc, dc):
        np.random.seed(seed)
        outputs_with_grads = [0]
        ref = gru_unit
        ref = partial(ref)


            

Reported by Pylint.

                      ref = gru_unit
        ref = partial(ref)

        t, n, d = input_tensor.shape
        assert d % 3 == 0
        d = d // 3
        ref = partial(ref, drop_states=drop_states,
                      sequence_lengths=sequence_lengths)


            

Reported by Pylint.

                                         input_tensor,
                           device_option=gc)
        print(str(net.Proto()))
        op = net._net.op[-1]
        inputs = [workspace.FetchBlob(name) for name in op.input]

        self.assertReferenceChecks(
            gc,
            op,

            

Reported by Pylint.

                                        **kwargs)

    def gru_base(self, create_rnn, ref, outputs_with_grads,
                 input_tensor, fwd_only, drop_states, linear_before_reset, gc, dc):

        print("GRU test parameters: ", locals())
        t, n, d = input_tensor.shape
        assert d % 3 == 0
        d = d // 3

            

Reported by Pylint.

              from typing import Optional, List
import torch
from torch.testing._internal.common_utils import TestCase, run_tests

# End-to-end tests of features in native_functions.yaml


class FloatListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[float]]):

            

Reported by Pylint.

              from typing import Optional, List
import torch
from torch.testing._internal.common_utils import TestCase, run_tests

# End-to-end tests of features in native_functions.yaml


class FloatListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[float]]):

            

Reported by Pylint.

              
class FloatListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[float]]):
        return torch._C._nn._test_optional_floatlist(values, incr)


class IntListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[int]]):
        return torch._C._nn._test_optional_intlist(values, incr)

            

Reported by Pylint.

              
class FloatListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[float]]):
        return torch._C._nn._test_optional_floatlist(values, incr)


class IntListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[int]]):
        return torch._C._nn._test_optional_intlist(values, incr)

            

Reported by Pylint.

              
class FloatListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[float]]):
        return torch._C._nn._test_optional_floatlist(values, incr)


class IntListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[int]]):
        return torch._C._nn._test_optional_intlist(values, incr)

            

Reported by Pylint.

              
class IntListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[int]]):
        return torch._C._nn._test_optional_intlist(values, incr)


class TestNativeFunctions(TestCase):

    #

            

Reported by Pylint.

              
class IntListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[int]]):
        return torch._C._nn._test_optional_intlist(values, incr)


class TestNativeFunctions(TestCase):

    #

            

Reported by Pylint.

              
class IntListWrapperModule(torch.nn.Module):
    def forward(self, values, incr: Optional[List[int]]):
        return torch._C._nn._test_optional_intlist(values, incr)


class TestNativeFunctions(TestCase):

    #

            

Reported by Pylint.

              
    def trace_optional_floatlist(self, const):
        def wrapper(values):
            return torch._C._nn._test_optional_floatlist(values, const)
        return torch.jit.trace(wrapper, torch.tensor([1.5, 2.5], dtype=torch.float))

    def test_optional_floatlist(self):
        self.do_test_optional_floatlist_with_module(FloatListWrapperModule())
        self.do_test_optional_floatlist_with_module(torch.jit.script(FloatListWrapperModule()))

            

Reported by Pylint.

              
    def trace_optional_floatlist(self, const):
        def wrapper(values):
            return torch._C._nn._test_optional_floatlist(values, const)
        return torch.jit.trace(wrapper, torch.tensor([1.5, 2.5], dtype=torch.float))

    def test_optional_floatlist(self):
        self.do_test_optional_floatlist_with_module(FloatListWrapperModule())
        self.do_test_optional_floatlist_with_module(torch.jit.script(FloatListWrapperModule()))

            

Reported by Pylint.

              import torch


class TorchTensorEngine(object):
    def rand(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.rand(shape, device=device, dtype=dtype, requires_grad=requires_grad)

    def randn(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.randn(shape, device=device, dtype=dtype, requires_grad=requires_grad)

            

Reported by Pylint.

                  def cat(self, inputs, dim=0):
        return torch.cat(inputs, dim=dim)

    def clamp(self, data, min, max):
        return torch.clamp(data, min=min, max=max)

    def relu(self, data):
        return torch.nn.functional.relu(data)


            

Reported by Pylint.

                  def cat(self, inputs, dim=0):
        return torch.cat(inputs, dim=dim)

    def clamp(self, data, min, max):
        return torch.clamp(data, min=min, max=max)

    def relu(self, data):
        return torch.nn.functional.relu(data)


            

Reported by Pylint.

              import torch


class TorchTensorEngine(object):
    def rand(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.rand(shape, device=device, dtype=dtype, requires_grad=requires_grad)

    def randn(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.randn(shape, device=device, dtype=dtype, requires_grad=requires_grad)

            

Reported by Pylint.

              import torch


class TorchTensorEngine(object):
    def rand(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.rand(shape, device=device, dtype=dtype, requires_grad=requires_grad)

    def randn(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.randn(shape, device=device, dtype=dtype, requires_grad=requires_grad)

            

Reported by Pylint.

              import torch


class TorchTensorEngine(object):
    def rand(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.rand(shape, device=device, dtype=dtype, requires_grad=requires_grad)

    def randn(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.randn(shape, device=device, dtype=dtype, requires_grad=requires_grad)

            

Reported by Pylint.

              import torch


class TorchTensorEngine(object):
    def rand(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.rand(shape, device=device, dtype=dtype, requires_grad=requires_grad)

    def randn(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.randn(shape, device=device, dtype=dtype, requires_grad=requires_grad)

            

Reported by Pylint.

              

class TorchTensorEngine(object):
    def rand(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.rand(shape, device=device, dtype=dtype, requires_grad=requires_grad)

    def randn(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.randn(shape, device=device, dtype=dtype, requires_grad=requires_grad)


            

Reported by Pylint.

              

class TorchTensorEngine(object):
    def rand(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.rand(shape, device=device, dtype=dtype, requires_grad=requires_grad)

    def randn(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.randn(shape, device=device, dtype=dtype, requires_grad=requires_grad)


            

Reported by Pylint.

                  def rand(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.rand(shape, device=device, dtype=dtype, requires_grad=requires_grad)

    def randn(self, shape, device=None, dtype=None, requires_grad=False):
        return torch.randn(shape, device=device, dtype=dtype, requires_grad=requires_grad)

    def nchw_rand(self, shape, device=None, requires_grad=False):
        return self.rand(shape, device=device, requires_grad=requires_grad)


            

Reported by Pylint.

              

def thread_runner(idx, testobj):
    global SUCCESS_COUNT
    testobj.assertEquals(scope.CurrentNameScope(), "")
    testobj.assertEquals(scope.CurrentDeviceScope(), None)
    namescope = "namescope_{}".format(idx)
    dsc = core.DeviceOption(workspace.GpuDeviceType, idx)
    with scope.DeviceScope(dsc):

            

Reported by Pylint.

              class TestScope(unittest.TestCase):

    def testNamescopeBasic(self):
        self.assertEquals(scope.CurrentNameScope(), "")

        with scope.NameScope("test_scope"):
            self.assertEquals(scope.CurrentNameScope(), "test_scope/")

        self.assertEquals(scope.CurrentNameScope(), "")

            

Reported by Pylint.

                      self.assertEquals(scope.CurrentNameScope(), "")

        with scope.NameScope("test_scope"):
            self.assertEquals(scope.CurrentNameScope(), "test_scope/")

        self.assertEquals(scope.CurrentNameScope(), "")

    def testNamescopeAssertion(self):
        self.assertEquals(scope.CurrentNameScope(), "")

            

Reported by Pylint.

                      with scope.NameScope("test_scope"):
            self.assertEquals(scope.CurrentNameScope(), "test_scope/")

        self.assertEquals(scope.CurrentNameScope(), "")

    def testNamescopeAssertion(self):
        self.assertEquals(scope.CurrentNameScope(), "")

        try:

            

Reported by Pylint.

                      self.assertEquals(scope.CurrentNameScope(), "")

    def testNamescopeAssertion(self):
        self.assertEquals(scope.CurrentNameScope(), "")

        try:
            with scope.NameScope("test_scope"):
                self.assertEquals(scope.CurrentNameScope(), "test_scope/")
                raise Exception()

            

Reported by Pylint.

              
        try:
            with scope.NameScope("test_scope"):
                self.assertEquals(scope.CurrentNameScope(), "test_scope/")
                raise Exception()
        except Exception:
            pass

        self.assertEquals(scope.CurrentNameScope(), "")

            

Reported by Pylint.

                          with scope.NameScope("test_scope"):
                self.assertEquals(scope.CurrentNameScope(), "test_scope/")
                raise Exception()
        except Exception:
            pass

        self.assertEquals(scope.CurrentNameScope(), "")

    def testEmptyNamescopeBasic(self):

            

Reported by Pylint.

                      except Exception:
            pass

        self.assertEquals(scope.CurrentNameScope(), "")

    def testEmptyNamescopeBasic(self):
        self.assertEquals(scope.CurrentNameScope(), "")

        with scope.NameScope("test_scope"):

            

Reported by Pylint.

                      self.assertEquals(scope.CurrentNameScope(), "")

    def testEmptyNamescopeBasic(self):
        self.assertEquals(scope.CurrentNameScope(), "")

        with scope.NameScope("test_scope"):
            with scope.EmptyNameScope():
                self.assertEquals(scope.CurrentNameScope(), "")
            self.assertEquals(scope.CurrentNameScope(), "test_scope/")

            

Reported by Pylint.

              
        with scope.NameScope("test_scope"):
            with scope.EmptyNameScope():
                self.assertEquals(scope.CurrentNameScope(), "")
            self.assertEquals(scope.CurrentNameScope(), "test_scope/")

    def testDevicescopeBasic(self):
        self.assertEquals(scope.CurrentDeviceScope(), None)


            

Reported by Pylint.

              import os
import re

from tensorboard.compat.proto.graph_pb2 import GraphDef
from tensorboard.compat.proto.node_def_pb2 import NodeDef
from tensorboard.compat.proto.tensor_shape_pb2 import TensorShapeProto

from builtins import bytes
from caffe2.proto import caffe2_pb2

            

Reported by Pylint.

              import re

from tensorboard.compat.proto.graph_pb2 import GraphDef
from tensorboard.compat.proto.node_def_pb2 import NodeDef
from tensorboard.compat.proto.tensor_shape_pb2 import TensorShapeProto

from builtins import bytes
from caffe2.proto import caffe2_pb2
from caffe2.python import core, workspace

            

Reported by Pylint.

              
from tensorboard.compat.proto.graph_pb2 import GraphDef
from tensorboard.compat.proto.node_def_pb2 import NodeDef
from tensorboard.compat.proto.tensor_shape_pb2 import TensorShapeProto

from builtins import bytes
from caffe2.proto import caffe2_pb2
from caffe2.python import core, workspace


            

Reported by Pylint.

                      # We don't care about the types, only the shapes
        shapes, _ = workspace.InferShapesAndTypes(nets)
        return shapes
    except Exception as e:
        logging.warning('Failed to compute shapes: %s', e)
        return {}


def model_to_graph_def(model, **kwargs):

            

Reported by Pylint.

              import copy
import logging
import os
import re

from tensorboard.compat.proto.graph_pb2 import GraphDef
from tensorboard.compat.proto.node_def_pb2 import NodeDef
from tensorboard.compat.proto.tensor_shape_pb2 import TensorShapeProto


            

Reported by Pylint.

              from tensorboard.compat.proto.node_def_pb2 import NodeDef
from tensorboard.compat.proto.tensor_shape_pb2 import TensorShapeProto

from builtins import bytes
from caffe2.proto import caffe2_pb2
from caffe2.python import core, workspace

from typing import Set, Dict, Tuple, List


            

Reported by Pylint.

              from tensorboard.compat.proto.tensor_shape_pb2 import TensorShapeProto

from builtins import bytes
from caffe2.proto import caffe2_pb2
from caffe2.python import core, workspace

from typing import Set, Dict, Tuple, List



            

Reported by Pylint.

              
from builtins import bytes
from caffe2.proto import caffe2_pb2
from caffe2.python import core, workspace

from typing import Set, Dict, Tuple, List


def _make_unique_name(seen: Set[str], name: str, min_version: int = 0):

            

Reported by Pylint.

              from caffe2.proto import caffe2_pb2
from caffe2.python import core, workspace

from typing import Set, Dict, Tuple, List


def _make_unique_name(seen: Set[str], name: str, min_version: int = 0):
    '''
    Make the name unique by appending a unique number to the name. Used for SSA.

            

Reported by Pylint.

                  Returns:
        x (string): A version of name that is not in seen.
    '''
    assert name is not None
    i = min_version
    x = '%s_%d' % (name, i) if i else name
    while x in seen:
        i += 1
        x = '%s_%d' % (name, i)

            

Reported by Bandit.

                          # TODO: Remove this once ScriptModule supports registering None buffer
            self.register_buffer(
                'bias',
                torch.zeros([dense_module.weight.size(0)], dtype=torch.float).to_mkldnn())

    @torch.jit.script_method
    def __getstate__(self):
        return (self.weight.to_dense(), self.bias.to_dense(), self.training)


            

Reported by Pylint.

                          # TODO: Remove this once ScriptModule supports registering None buffer
            self.register_buffer(
                'bias',
                torch.zeros([dense_module.weight.size(0)], dtype=torch.float).to_mkldnn())

    @torch.jit.script_method
    def __getstate__(self):
        return (self.weight.to_dense(), self.bias.to_dense(), self.training)


            

Reported by Pylint.

                          # TODO: Remove this once ScriptModule supports registering None buffer
            self.register_buffer(
                'bias',
                torch.zeros([dense_module.weight.size(0)], dtype=torch.float).to_mkldnn())

    @torch.jit.script_method
    def __getstate__(self):
        return (self.weight.to_dense(), self.bias.to_dense(), self.training)


            

Reported by Pylint.

                          # TODO: Remove this once ScriptModule supports registering None buffer
            self.register_buffer(
                'bias',
                torch.zeros([dense_module.weight.size(0)], dtype=torch.float).to_mkldnn())

    @torch.jit.script_method
    def __getstate__(self):
        return (self.weight.to_dense(), self.bias.to_dense(), self.training)


            

Reported by Pylint.

              
    @torch.jit.script_method
    def forward(self, x):
        return torch.mkldnn_convolution(
            x,
            self.weight,
            self.bias,
            self.padding,
            self.stride,

            

Reported by Pylint.

              
    @torch.jit.script_method
    def forward(self, x):
        return torch.batch_norm(
            x,
            self.weight,
            self.bias,
            self.running_mean,
            self.running_var,

            

Reported by Pylint.

                      )


def to_mkldnn(module, dtype=torch.float):
    assert dtype in [torch.float, torch.bfloat16], \
        "MKLDNN only support float or bfloat16 path now"

    def m_fn(m, d):
        if isinstance(m, torch.nn.Linear):

            

Reported by Pylint.

              

def to_mkldnn(module, dtype=torch.float):
    assert dtype in [torch.float, torch.bfloat16], \
        "MKLDNN only support float or bfloat16 path now"

    def m_fn(m, d):
        if isinstance(m, torch.nn.Linear):
            return MkldnnLinear(m, d)

            

Reported by Pylint.

              

def to_mkldnn(module, dtype=torch.float):
    assert dtype in [torch.float, torch.bfloat16], \
        "MKLDNN only support float or bfloat16 path now"

    def m_fn(m, d):
        if isinstance(m, torch.nn.Linear):
            return MkldnnLinear(m, d)

            

Reported by Pylint.

                          # we use fp32 dtype.
            self.register_buffer('bias', dense_module.bias.to_mkldnn())
        else:
            # TODO: Remove this once ScriptModule supports registering None buffer
            self.register_buffer(
                'bias',
                torch.zeros([dense_module.weight.size(0)], dtype=torch.float).to_mkldnn())

    @torch.jit.script_method

            

Reported by Pylint.

              import torch
from ..optimizer import Optimizer
from collections import defaultdict

class RMSprop(Optimizer):
    r"""Implements RMSprop algorithm.

    Proposed by G. Hinton in his
    `course <https://www.cs.toronto.edu/~tijmen/csc321/slides/lecture_slides_lec6.pdf>`_.

            

Reported by Pylint.

                                  # State initialization
                    if len(state) == 0:
                        state['step'] = 0
                        state['square_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['momentum'] > 0:
                            state['momentum_buffer'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['centered']:
                            state['grad_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)


            

Reported by Pylint.

                                  # State initialization
                    if len(state) == 0:
                        state['step'] = 0
                        state['square_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['momentum'] > 0:
                            state['momentum_buffer'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['centered']:
                            state['grad_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)


            

Reported by Pylint.

                                      state['step'] = 0
                        state['square_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['momentum'] > 0:
                            state['momentum_buffer'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['centered']:
                            state['grad_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)

                        state['step'] += 1


            

Reported by Pylint.

                                      state['step'] = 0
                        state['square_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['momentum'] > 0:
                            state['momentum_buffer'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['centered']:
                            state['grad_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)

                        state['step'] += 1


            

Reported by Pylint.

                                      if group['momentum'] > 0:
                            state['momentum_buffer'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['centered']:
                            state['grad_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)

                        state['step'] += 1

                    states.append(state)
                    square_avg.append(state['square_avg'])

            

Reported by Pylint.

                                      if group['momentum'] > 0:
                            state['momentum_buffer'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                        if group['centered']:
                            state['grad_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)

                        state['step'] += 1

                    states.append(state)
                    square_avg.append(state['square_avg'])

            

Reported by Pylint.

                                  square_avg.append(state['square_avg'])

            if group['weight_decay'] != 0:
                torch._foreach_add_(grads, params_with_grad, alpha=group['weight_decay'])

            torch._foreach_mul_(square_avg, alpha)
            torch._foreach_addcmul_(square_avg, grads, grads, value=1 - alpha)

            if group['centered']:

            

Reported by Pylint.

                          if group['weight_decay'] != 0:
                torch._foreach_add_(grads, params_with_grad, alpha=group['weight_decay'])

            torch._foreach_mul_(square_avg, alpha)
            torch._foreach_addcmul_(square_avg, grads, grads, value=1 - alpha)

            if group['centered']:
                grad_avgs = [s['grad_avg'] for s in states]
                torch._foreach_mul_(grad_avgs, alpha)

            

Reported by Pylint.

                              torch._foreach_add_(grads, params_with_grad, alpha=group['weight_decay'])

            torch._foreach_mul_(square_avg, alpha)
            torch._foreach_addcmul_(square_avg, grads, grads, value=1 - alpha)

            if group['centered']:
                grad_avgs = [s['grad_avg'] for s in states]
                torch._foreach_mul_(grad_avgs, alpha)
                torch._foreach_add_(grad_avgs, grads, alpha=1 - alpha)

            

Reported by Pylint.

                      def myhelper(model, val=-1):
            return val

        if not brew.has_helper(myhelper):
            brew.Register(myhelper)
        self.myhelper = myhelper

        def myhelper2(model, val=-1):
            return val

            

Reported by Pylint.

                          return val

        if not brew.has_helper(myhelper):
            brew.Register(myhelper)
        self.myhelper = myhelper

        def myhelper2(model, val=-1):
            return val


            

Reported by Pylint.

                      def myhelper2(model, val=-1):
            return val

        if not brew.has_helper(myhelper2):
            brew.Register(myhelper2)
        self.myhelper2 = myhelper2
        self.model = ModelHelper(name="test_model")

    def test_dropout(self):

            

Reported by Pylint.

                          return val

        if not brew.has_helper(myhelper2):
            brew.Register(myhelper2)
        self.myhelper2 = myhelper2
        self.model = ModelHelper(name="test_model")

    def test_dropout(self):
        p = 0.2

            

Reported by Pylint.

                      myhelper2 = self.myhelper2
        n = 15
        with brew.arg_scope([myhelper], val=n):
            res = brew.myhelper(self.model)
        self.assertEqual(n, res)

        with brew.arg_scope([myhelper, myhelper2], val=n):
            res1 = brew.myhelper(self.model)
            res2 = brew.myhelper2(self.model)

            

Reported by Pylint.

                      self.assertEqual(n, res)

        with brew.arg_scope([myhelper, myhelper2], val=n):
            res1 = brew.myhelper(self.model)
            res2 = brew.myhelper2(self.model)
        self.assertEqual([n, n], [res1, res2])

    def test_arg_scope_single(self):
        X = np.random.rand(64, 3, 32, 32).astype(np.float32) - 0.5

            

Reported by Pylint.

              
        with brew.arg_scope([myhelper, myhelper2], val=n):
            res1 = brew.myhelper(self.model)
            res2 = brew.myhelper2(self.model)
        self.assertEqual([n, n], [res1, res2])

    def test_arg_scope_single(self):
        X = np.random.rand(64, 3, 32, 32).astype(np.float32) - 0.5


            

Reported by Pylint.

                      with brew.arg_scope([myhelper], val=-3), \
                brew.arg_scope([myhelper], val=-2):
            with brew.arg_scope([myhelper], val=n):
                res = brew.myhelper(self.model)
                self.assertEqual(n, res)
            res = brew.myhelper(self.model)
            self.assertEqual(res, -2)

        res = brew.myhelper(self.model, val=15)

            

Reported by Pylint.

                          with brew.arg_scope([myhelper], val=n):
                res = brew.myhelper(self.model)
                self.assertEqual(n, res)
            res = brew.myhelper(self.model)
            self.assertEqual(res, -2)

        res = brew.myhelper(self.model, val=15)
        self.model.Validate()
        self.assertEqual(res, 15)

            

Reported by Pylint.

                          res = brew.myhelper(self.model)
            self.assertEqual(res, -2)

        res = brew.myhelper(self.model, val=15)
        self.model.Validate()
        self.assertEqual(res, 15)

    def test_double_register(self):
        myhelper = self.myhelper

            

Reported by Pylint.

              import numpy as np
import unittest

import caffe2.python.fakelowp.init_shared_libs  # noqa
from hypothesis import given, settings
from hypothesis import strategies as st
from caffe2.proto import caffe2_pb2
from caffe2.python import core
from caffe2.python import workspace

            

Reported by Pylint.

              import unittest

import caffe2.python.fakelowp.init_shared_libs  # noqa
from hypothesis import given, settings
from hypothesis import strategies as st
from caffe2.proto import caffe2_pb2
from caffe2.python import core
from caffe2.python import workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net

            

Reported by Pylint.

              
import caffe2.python.fakelowp.init_shared_libs  # noqa
from hypothesis import given, settings
from hypothesis import strategies as st
from caffe2.proto import caffe2_pb2
from caffe2.python import core
from caffe2.python import workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info

            

Reported by Pylint.

              import caffe2.python.fakelowp.init_shared_libs  # noqa
from hypothesis import given, settings
from hypothesis import strategies as st
from caffe2.proto import caffe2_pb2
from caffe2.python import core
from caffe2.python import workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import datetime

            

Reported by Pylint.

              from hypothesis import given, settings
from hypothesis import strategies as st
from caffe2.proto import caffe2_pb2
from caffe2.python import core
from caffe2.python import workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import datetime
import caffe2.python.serialized_test.serialized_test_util as serial

            

Reported by Pylint.

              from hypothesis import strategies as st
from caffe2.proto import caffe2_pb2
from caffe2.python import core
from caffe2.python import workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import datetime
import caffe2.python.serialized_test.serialized_test_util as serial


            

Reported by Pylint.

              from caffe2.proto import caffe2_pb2
from caffe2.python import core
from caffe2.python import workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import datetime
import caffe2.python.serialized_test.serialized_test_util as serial

core.GlobalInit(["caffe2", "--caffe2_log_level=-3", "--glow_global_fp16=1"])

            

Reported by Pylint.

              from caffe2.python import core
from caffe2.python import workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import datetime
import caffe2.python.serialized_test.serialized_test_util as serial

core.GlobalInit(["caffe2", "--caffe2_log_level=-3", "--glow_global_fp16=1"])


            

Reported by Pylint.

              from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import datetime
import caffe2.python.serialized_test.serialized_test_util as serial

core.GlobalInit(["caffe2", "--caffe2_log_level=-3", "--glow_global_fp16=1"])

GLOW_MATMUL_RTOL = 0


            

Reported by Pylint.

              import numpy as np
import unittest

import caffe2.python.fakelowp.init_shared_libs  # noqa
from hypothesis import given, settings
from hypothesis import strategies as st
from caffe2.proto import caffe2_pb2
from caffe2.python import core
from caffe2.python import workspace

            

Reported by Pylint.