import torch
from torch import Tensor
from .module import Module
from ..parameter import Parameter
from ..utils.rnn import PackedSequence
from .. import init
from ... import _VF


            

Reported by Pylint.

              import torch
from torch import Tensor
from .module import Module
from ..parameter import Parameter
from ..utils.rnn import PackedSequence
from .. import init
from ... import _VF

_rnn_impls = {

            

Reported by Pylint.

              from torch import Tensor
from .module import Module
from ..parameter import Parameter
from ..utils.rnn import PackedSequence
from .. import init
from ... import _VF

_rnn_impls = {
    'RNN_TANH': _VF.rnn_tanh,

            

Reported by Pylint.

              from .module import Module
from ..parameter import Parameter
from ..utils.rnn import PackedSequence
from .. import init
from ... import _VF

_rnn_impls = {
    'RNN_TANH': _VF.rnn_tanh,
    'RNN_RELU': _VF.rnn_relu,

            

Reported by Pylint.

              from ..parameter import Parameter
from ..utils.rnn import PackedSequence
from .. import init
from ... import _VF

_rnn_impls = {
    'RNN_TANH': _VF.rnn_tanh,
    'RNN_RELU': _VF.rnn_relu,
}

            

Reported by Pylint.

                              real_hidden_size = proj_size if proj_size > 0 else hidden_size
                layer_input_size = input_size if layer == 0 else real_hidden_size * num_directions

                w_ih = Parameter(torch.empty((gate_size, layer_input_size), **factory_kwargs))
                w_hh = Parameter(torch.empty((gate_size, real_hidden_size), **factory_kwargs))
                b_ih = Parameter(torch.empty(gate_size, **factory_kwargs))
                # Second bias vector included for CuDNN compatibility. Only one
                # bias vector is needed in standard definition.
                b_hh = Parameter(torch.empty(gate_size, **factory_kwargs))

            

Reported by Pylint.

                              layer_input_size = input_size if layer == 0 else real_hidden_size * num_directions

                w_ih = Parameter(torch.empty((gate_size, layer_input_size), **factory_kwargs))
                w_hh = Parameter(torch.empty((gate_size, real_hidden_size), **factory_kwargs))
                b_ih = Parameter(torch.empty(gate_size, **factory_kwargs))
                # Second bias vector included for CuDNN compatibility. Only one
                # bias vector is needed in standard definition.
                b_hh = Parameter(torch.empty(gate_size, **factory_kwargs))
                layer_params: Tuple[Tensor, ...] = ()

            

Reported by Pylint.

              
                w_ih = Parameter(torch.empty((gate_size, layer_input_size), **factory_kwargs))
                w_hh = Parameter(torch.empty((gate_size, real_hidden_size), **factory_kwargs))
                b_ih = Parameter(torch.empty(gate_size, **factory_kwargs))
                # Second bias vector included for CuDNN compatibility. Only one
                # bias vector is needed in standard definition.
                b_hh = Parameter(torch.empty(gate_size, **factory_kwargs))
                layer_params: Tuple[Tensor, ...] = ()
                if self.proj_size == 0:

            

Reported by Pylint.

                              b_ih = Parameter(torch.empty(gate_size, **factory_kwargs))
                # Second bias vector included for CuDNN compatibility. Only one
                # bias vector is needed in standard definition.
                b_hh = Parameter(torch.empty(gate_size, **factory_kwargs))
                layer_params: Tuple[Tensor, ...] = ()
                if self.proj_size == 0:
                    if bias:
                        layer_params = (w_ih, w_hh, b_ih, b_hh)
                    else:

            

Reported by Pylint.

                                  else:
                        layer_params = (w_ih, w_hh)
                else:
                    w_hr = Parameter(torch.empty((proj_size, hidden_size), **factory_kwargs))
                    if bias:
                        layer_params = (w_ih, w_hh, b_ih, b_hh, w_hr)
                    else:
                        layer_params = (w_ih, w_hh, w_hr)


            

Reported by Pylint.

              # TODO(jiayq): as more and more tests are moving to hypothesis test, we
# can gradually remove this test script. DO NOT ADD MORE TESTS TO THIS
# FILE.





import numpy as np

            

Reported by Pylint.

                          res = device_checker.CheckSimple(op, [X], [0])
            self.assertTrue(res)
            for checker in gradient_checkers:
                res, grad, grad_estimated = checker.CheckSimple(op, [X], 0, [0])
                self.assertTrue(res)


class TestFlatten(test_util.TestCase):


            

Reported by Pylint.

                          res = device_checker.CheckSimple(op, [X], [0])
            self.assertTrue(res)
            for checker in gradient_checkers:
                res, grad, grad_estimated = checker.CheckSimple(op, [X], 0, [0])
                self.assertTrue(res)


class TestFlatten(test_util.TestCase):


            

Reported by Pylint.

                      res = device_checker.CheckSimple(op, [X], [0])
        self.assertTrue(res)
        for checker in gradient_checkers:
            res, grad, grad_estimated = checker.CheckSimple(op, [X], 0, [0])
            self.assertTrue(res)


class TestConcat(test_util.TestCase):


            

Reported by Pylint.

                      res = device_checker.CheckSimple(op, [X], [0])
        self.assertTrue(res)
        for checker in gradient_checkers:
            res, grad, grad_estimated = checker.CheckSimple(op, [X], 0, [0])
            self.assertTrue(res)


class TestConcat(test_util.TestCase):


            

Reported by Pylint.

                              res = device_checker.CheckSimple(op, Xs, [0])
                self.assertTrue(res)
                for checker in gradient_checkers:
                    res, grad, grad_estimated = checker.CheckSimple(op, Xs, i,
                                                                    [0])
                    self.assertTrue(res)

    def testConcatNCHW(self):
        for input_size, d1, d2, d3, d4 in self.test_configs:

            

Reported by Pylint.

                              res = device_checker.CheckSimple(op, Xs, [0])
                self.assertTrue(res)
                for checker in gradient_checkers:
                    res, grad, grad_estimated = checker.CheckSimple(op, Xs, i,
                                                                    [0])
                    self.assertTrue(res)

    def testConcatNCHW(self):
        for input_size, d1, d2, d3, d4 in self.test_configs:

            

Reported by Pylint.

                              res = device_checker.CheckSimple(op, Xs, [0])
                self.assertTrue(res)
                for checker in gradient_checkers:
                    res, grad, grad_estimated = checker.CheckSimple(op, Xs, i,
                                                                    [0])
                    self.assertTrue(res)


class TestRelu(test_util.TestCase):

            

Reported by Pylint.

                              res = device_checker.CheckSimple(op, Xs, [0])
                self.assertTrue(res)
                for checker in gradient_checkers:
                    res, grad, grad_estimated = checker.CheckSimple(op, Xs, i,
                                                                    [0])
                    self.assertTrue(res)


class TestRelu(test_util.TestCase):

            

Reported by Pylint.

                          res = device_checker.CheckSimple(op, [X], [0])
            self.assertTrue(res)
            for checker in gradient_checkers:
                res, grad, grad_estimated = checker.CheckSimple(op, [X], 0, [0])
                self.assertTrue(res)


class TestTanh(test_util.TestCase):


            

Reported by Pylint.

              import math
import torch
import torch.nn as nn
import torch.backends.mkldnn
from torch.nn import Conv2d, BatchNorm2d, ReLU, init
from torch.nn.intrinsic.qat import ConvBn2d, ConvBnReLU2d
from torch.nn.modules.utils import _pair
import torch.nn.quantized as nnq
from torch.quantization import (

            

Reported by Pylint.

              import math
import torch
import torch.nn as nn
import torch.backends.mkldnn
from torch.nn import Conv2d, BatchNorm2d, ReLU, init
from torch.nn.intrinsic.qat import ConvBn2d, ConvBnReLU2d
from torch.nn.modules.utils import _pair
import torch.nn.quantized as nnq
from torch.quantization import (

            

Reported by Pylint.

              import math
import torch
import torch.nn as nn
import torch.backends.mkldnn
from torch.nn import Conv2d, BatchNorm2d, ReLU, init
from torch.nn.intrinsic.qat import ConvBn2d, ConvBnReLU2d
from torch.nn.modules.utils import _pair
import torch.nn.quantized as nnq
from torch.quantization import (

            

Reported by Pylint.

              import torch
import torch.nn as nn
import torch.backends.mkldnn
from torch.nn import Conv2d, BatchNorm2d, ReLU, init
from torch.nn.intrinsic.qat import ConvBn2d, ConvBnReLU2d
from torch.nn.modules.utils import _pair
import torch.nn.quantized as nnq
from torch.quantization import (
    prepare,

            

Reported by Pylint.

              import torch.nn as nn
import torch.backends.mkldnn
from torch.nn import Conv2d, BatchNorm2d, ReLU, init
from torch.nn.intrinsic.qat import ConvBn2d, ConvBnReLU2d
from torch.nn.modules.utils import _pair
import torch.nn.quantized as nnq
from torch.quantization import (
    prepare,
    convert,

            

Reported by Pylint.

              import torch.backends.mkldnn
from torch.nn import Conv2d, BatchNorm2d, ReLU, init
from torch.nn.intrinsic.qat import ConvBn2d, ConvBnReLU2d
from torch.nn.modules.utils import _pair
import torch.nn.quantized as nnq
from torch.quantization import (
    prepare,
    convert,
    prepare_qat,

            

Reported by Pylint.

              from torch.nn import Conv2d, BatchNorm2d, ReLU, init
from torch.nn.intrinsic.qat import ConvBn2d, ConvBnReLU2d
from torch.nn.modules.utils import _pair
import torch.nn.quantized as nnq
from torch.quantization import (
    prepare,
    convert,
    prepare_qat,
    quantize_qat,

            

Reported by Pylint.

              from torch.nn.intrinsic.qat import ConvBn2d, ConvBnReLU2d
from torch.nn.modules.utils import _pair
import torch.nn.quantized as nnq
from torch.quantization import (
    prepare,
    convert,
    prepare_qat,
    quantize_qat,
    QuantStub,

            

Reported by Pylint.

                  default_qat_qconfig,
    FixedQParamsFakeQuantize,
)
from torch.testing._internal.common_utils import TestCase

from torch.testing._internal.common_quantization import (
    QuantizationTestCase,
    QuantStubModel,
    ManualLinearQATModel,

            

Reported by Pylint.

              )
from torch.testing._internal.common_utils import TestCase

from torch.testing._internal.common_quantization import (
    QuantizationTestCase,
    QuantStubModel,
    ManualLinearQATModel,
    ManualConvLinearQATModel,
    TwoLayerLinearModel,

            

Reported by Pylint.

              from typing import Dict, Optional

import numpy as np
import torch
from torch import nn
from torch.testing._internal.common_utils import TestCase, run_tests


class StaticModule:

            

Reported by Pylint.

              
import numpy as np
import torch
from torch import nn
from torch.testing._internal.common_utils import TestCase, run_tests


class StaticModule:
    def __init__(self, scripted):

            

Reported by Pylint.

              import numpy as np
import torch
from torch import nn
from torch.testing._internal.common_utils import TestCase, run_tests


class StaticModule:
    def __init__(self, scripted):
        # this is an nn.Module

            

Reported by Pylint.

              

def linear_shim(
    input: torch.Tensor, weight: torch.Tensor, bias: Optional[torch.Tensor] = None
) -> torch.Tensor:
    output = input.matmul(weight.t())
    if bias is not None:
        output += bias
    ret = output

            

Reported by Pylint.

              

class MultiHeadAttentionLayer(nn.Module):
    def __init__(self, hid_dim, n_heads, dropout, device):
        super().__init__()
        assert hid_dim % n_heads == 0
        self.hid_dim = hid_dim
        self.n_heads = n_heads
        self.head_dim = hid_dim // n_heads

            

Reported by Pylint.

                      # self.dropout = nn.Dropout(dropout)
        self.scale = torch.sqrt(torch.FloatTensor([self.head_dim])).to(device)

    def forward(self, query, key, value, mask):
        batch_size = query.shape[0]
        Q = self.fc_q(query)
        K = self.fc_k(key)
        V = self.fc_v(value)
        Q = Q.view(batch_size, -1, self.n_heads, self.head_dim).permute(0, 2, 1, 3)

            

Reported by Pylint.

              
def loop_graph(a, b, iters: int):
    c = a + b * 2
    for i in range(iters):
        c = c + b
        c *= 2
        c -= a
    return c


            

Reported by Pylint.

                      HID_DIM = 256
        QUERY_LEN = 8
        BATCH_SIZE = 128
        LAYERS = 3
        HEADS = 8
        DROPOUT = 0.1
        device = torch.device("cpu")
        attention = MultiHeadAttentionLayer(HID_DIM, HEADS, DROPOUT, device).to(device)
        with torch.no_grad():

            

Reported by Pylint.

                      HID_DIM = 256
        QUERY_LEN = 8
        BATCH_SIZE = 128
        LAYERS = 3
        HEADS = 8
        DROPOUT = 0.1
        device = torch.device("cpu")
        attention = MultiHeadAttentionLayer(HID_DIM, HEADS, DROPOUT, device).to(device)
        with torch.no_grad():

            

Reported by Pylint.

                      attention_a = StaticModule(attention)

        attention_a.benchmark([src, src, src, src_mask], {}, 2, 2)
        metrics = attention_a.benchmark_individual_ops(
            [src, src, src, src_mask], {}, 2, 2
        )

    def test_mlp(self):
        # Arguments taken from benchmark script, ./bench/dlrm_s_benchmark.sh

            

Reported by Pylint.

              
from typing import Any, List

import torch
from torch.testing._internal.jit_utils import JitTestCase, make_global


# Make the helper files in test/ importable
pytorch_test_dir = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))

            

Reported by Pylint.

              from typing import Any, List

import torch
from torch.testing._internal.jit_utils import JitTestCase, make_global


# Make the helper files in test/ importable
pytorch_test_dir = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
sys.path.append(pytorch_test_dir)

            

Reported by Pylint.

                          def __init__(self):
                self.count = 1

            def __enter__(self, incr: int):
                self.count += incr

            def __exit__(self, type: Any, value: Any, tb: Any):
                pass


            

Reported by Pylint.

                          def __enter__(self):
                self.count += 1

            def __exit__(self, type: Any, value: Any):
                pass

        @torch.jit.script
        class ExitIncorrectTypes(object):
            """

            

Reported by Pylint.

                          return x

        def test_enter_without_object():
            with "not_object" as obj:
                pass

        test_tensor = torch.randn(5, dtype=torch.double)

        with self.assertRaisesRegexWithHighlight(

            

Reported by Pylint.

                              self.count.add_(0.3)
                return self.count

            def __exit__(self, type: Any, value: Any, tb: Any) -> bool:
                self.count.sub_(0.3)
                return True

        make_global(Context)


            

Reported by Pylint.

                          """
            c = Context(1)

            with c as mult:
                pass

            x *= c.count
            return x


            

Reported by Pylint.

                          x = y + y
            return x

        def test_conditional_early_return(x: torch.Tensor, c: Context) -> torch.Tensor:
            """
            Test that conditionally returning early from inside a with-statement works
            as expected.
            """
            with c as mult:

            

Reported by Pylint.

                              self.count.add_(0.3)
                return self.count

            def __exit__(self, type: Any, value: Any, tb: Any):
                self.count.sub_(0.3)

        make_global(Context)

        def test_basic(x: torch.Tensor) -> torch.Tensor:

            

Reported by Pylint.

                          x = y + y
            return x

        def test_conditional_early_return(x: torch.Tensor, c: Context) -> torch.Tensor:
            """
            Test that conditionally returning early from inside a with-statement works
            as expected.
            """
            with c:

            

Reported by Pylint.

              import functools

import torch
from ..parameter import Parameter
import torch.utils.hooks as hooks

from torch import Tensor, device, dtype
from typing import Union, Tuple, Any, Callable, Iterator, Set, Optional, overload, TypeVar, Mapping, Dict, List
from ...utils.hooks import RemovableHandle

            

Reported by Pylint.

              from ..parameter import Parameter
import torch.utils.hooks as hooks

from torch import Tensor, device, dtype
from typing import Union, Tuple, Any, Callable, Iterator, Set, Optional, overload, TypeVar, Mapping, Dict, List
from ...utils.hooks import RemovableHandle

_grad_t = Union[Tuple[Tensor, ...], Tensor]
# See https://mypy.readthedocs.io/en/latest/generics.html#generic-methods-and-generic-self for the use

            

Reported by Pylint.

              from ..parameter import Parameter
import torch.utils.hooks as hooks

from torch import Tensor, device, dtype
from typing import Union, Tuple, Any, Callable, Iterator, Set, Optional, overload, TypeVar, Mapping, Dict, List
from ...utils.hooks import RemovableHandle

_grad_t = Union[Tuple[Tensor, ...], Tensor]
# See https://mypy.readthedocs.io/en/latest/generics.html#generic-methods-and-generic-self for the use

            

Reported by Pylint.

              
from torch import Tensor, device, dtype
from typing import Union, Tuple, Any, Callable, Iterator, Set, Optional, overload, TypeVar, Mapping, Dict, List
from ...utils.hooks import RemovableHandle

_grad_t = Union[Tuple[Tensor, ...], Tensor]
# See https://mypy.readthedocs.io/en/latest/generics.html#generic-methods-and-generic-self for the use
# of `T` to annotate `self`. Many methods of `Module` return `self` and we want those return values to be
# the type of the subclass, not the looser type of `Module`.

            

Reported by Pylint.

                          module._apply(fn)

        def compute_should_use_set_data(tensor, tensor_applied):
            if torch._has_compatible_shallow_copy_type(tensor, tensor_applied):
                # If the new tensor has compatible tensor type as the existing tensor,
                # the current behavior is to change the tensor in-place using `.data =`,
                # and the future behavior is to overwrite the existing tensor. However,
                # changing the current behavior is a BC-breaking change, and we want it
                # to happen in future releases. So for now we introduce the

            

Reported by Pylint.

                      Returns:
            Module: self
        """
        return self._apply(lambda t: torch.empty_like(t, device=device))

    @overload
    def to(self: T, device: Optional[Union[int, device]] = ..., dtype: Optional[Union[dtype, str]] = ...,
           non_blocking: bool = ...) -> T:
        ...

            

Reported by Pylint.

                  return s


r"""This tracks hooks common to all modules that are executed before/after
calling forward and backward. This is global state used for debugging/profiling
purposes"""
_global_backward_hooks: Dict[int, Callable] = OrderedDict()
_global_is_full_backward_hook: Optional[bool] = None
_global_forward_pre_hooks: Dict[int, Callable] = OrderedDict()

            

Reported by Pylint.

                          ``handle.remove()``

    """
    global _global_is_full_backward_hook
    if _global_is_full_backward_hook is True:
        raise RuntimeError("Cannot use both regular backward hooks and full backward hooks as a "
                           "global Module hook. Please use only one of them.")

    _global_is_full_backward_hook = False

            

Reported by Pylint.

                          ``handle.remove()``

    """
    global _global_is_full_backward_hook
    if _global_is_full_backward_hook is False:
        raise RuntimeError("Cannot use both regular backward hooks and full backward hooks as a "
                           "global Module hook. Please use only one of them.")

    _global_is_full_backward_hook = True

            

Reported by Pylint.

              # Trick mypy into not applying contravariance rules to inputs by defining
# forward as a value, rather than a function.  See also
# https://github.com/python/mypy/issues/8795
def _forward_unimplemented(self, *input: Any) -> None:
    r"""Defines the computation performed at every call.

    Should be overridden by all subclasses.

    .. note::

            

Reported by Pylint.

              

from caffe2.python import core, workspace
from hypothesis import assume, given, settings
from caffe2.proto import caffe2_pb2
import caffe2.python.hypothesis_test_util as hu
import caffe2.python.serialized_test.serialized_test_util as serial
import hypothesis.strategies as st
import numpy as np

            

Reported by Pylint.

              from caffe2.proto import caffe2_pb2
import caffe2.python.hypothesis_test_util as hu
import caffe2.python.serialized_test.serialized_test_util as serial
import hypothesis.strategies as st
import numpy as np
import random


class TestUtilityOps(serial.SerializedTestCase):

            

Reported by Pylint.

                              "Slice", ["X", "starts", "ends"], ["Y"], device_option=gc
            )

            def slice_ref(x, starts, ends):
                slc = [slice(None)] * x.ndim
                slc[dim] = slice(slice_start, slice_end)
                return [x[slc]]
            inputs = [X, starts, ends]


            

Reported by Pylint.

                              "Slice", ["X", "starts", "ends"], ["Y"], device_option=gc
            )

            def slice_ref(x, starts, ends):
                slc = [slice(None)] * x.ndim
                slc[dim] = slice(slice_start, slice_end)
                return [x[slc]]
            inputs = [X, starts, ends]


            

Reported by Pylint.

                         engine=st.sampled_from(['CUDNN', None]),
           **hu.gcs)
    @settings(deadline=10000)
    def test_transpose(self, dtype, ndims, seed, null_axes, engine, gc, dc):
        if (gc.device_type == caffe2_pb2.CUDA and engine == "CUDNN"):
            # cudnn 5.1 does not support int.
            assume(workspace.GetCuDNNVersion() >= 6000 or dtype != np.int32)

        dims = (np.random.rand(ndims) * 16 + 1).astype(np.int32)

            

Reported by Pylint.

                  @given(m=st.integers(5, 10), n=st.integers(5, 10),
           o=st.integers(5, 10), nans=st.booleans(), **hu.gcs)
    @settings(deadline=10000)
    def test_nan_check(self, m, n, o, nans, gc, dc):
        other = np.array([1, 2, 3]).astype(np.float32)
        X = np.random.rand(m, n, o).astype(np.float32)
        if nans:
            x_nan = np.random.randint(0, m)
            y_nan = np.random.randint(0, n)

            

Reported by Pylint.

                      # print('nans: {}'.format(nans))
        # print(X)

        def nan_reference(X, Y):
            if not np.isnan(X).any():
                return [X]
            else:
                return [np.array([])]


            

Reported by Pylint.

                              reference=nan_reference,
            )
            if nans:
                self.assertTrue(False, "Did not fail when presented with NaN!")
        except RuntimeError:
            self.assertTrue(nans, "No NaNs but failed")

        try:
            self.assertGradientChecks(

            

Reported by Pylint.

                              outputs_with_grads=[0],
            )
            if nans:
                self.assertTrue(False, "Did not fail when gradient had NaN!")
        except RuntimeError:
            pass

    @serial.given(n=st.integers(4, 5), m=st.integers(6, 7),
           d=st.integers(2, 3), **hu.gcs)

            

Reported by Pylint.

                      ),
        **hu.gcs_cpu_only)
    @settings(deadline=10000)
    def test_lengths_gather(self, inputs, gc, dc):
        items = inputs[0]
        lengths = inputs[1]
        indices = inputs[2]

        def lengths_gather_op(items, lengths, indices):

            

Reported by Pylint.

                  """ A feature set has 2 types of features"""

    dense_features: torch.Tensor
    sparse_features: torch.LongTensor
    values: torch.Tensor


def _call_method(method, rref, *args, **kwargs):
    return method(rref.local_value(), *args, **kwargs)

            

Reported by Pylint.

                      self.em = nn.EmbeddingBag(
            num_embeddings,
            embedding_dim,
            _weight=torch.tensor([init_em] * num_embeddings),
        )

    def forward(self, input: torch.Tensor):
        gLogger.debug(f"Running RemoteEM.forward() on: {input}")
        return self.em(input, offsets=torch.LongTensor(range(input.shape[0])))

            

Reported by Pylint.

              
    def forward(self, input: torch.Tensor):
        gLogger.debug(f"Running RemoteEM.forward() on: {input}")
        return self.em(input, offsets=torch.LongTensor(range(input.shape[0])))


# Return a linear module with predefined parameters.
def getLinear(d_in, d_out):
    l = nn.Linear(d_in, d_out, bias=False)

            

Reported by Pylint.

              # Return a linear module with predefined parameters.
def getLinear(d_in, d_out):
    l = nn.Linear(d_in, d_out, bias=False)
    w = torch.ones((d_out, d_in))
    w[0][0] = -1
    w.requires_grad_()
    l.weight.data = w
    return l


            

Reported by Pylint.

                      # The same size of mini batch.
        assert sparse.shape[0] == input.dense_features.shape[0]
        dense = self.fc1(input.dense_features)
        x = torch.cat((dense, sparse), 1)
        gLogger.debug(f"Concatenated feature: {x}")
        x = _remote_method(RemoteNet.forward, self.remote_net_rref, x)
        return self.fc2(x)



            

Reported by Pylint.

              def get_training_examples():
    n = 16
    training_examples = FeatureSet(
        dense_features=torch.zeros((n, D_DENSE)),
        sparse_features=torch.zeros(n, dtype=torch.long),
        values=torch.zeros(n),
    )
    idx = 0
    # Every example has another one that has exactly the same features but an

            

Reported by Pylint.

                  n = 16
    training_examples = FeatureSet(
        dense_features=torch.zeros((n, D_DENSE)),
        sparse_features=torch.zeros(n, dtype=torch.long),
        values=torch.zeros(n),
    )
    idx = 0
    # Every example has another one that has exactly the same features but an
    # opposite value. Therefore, their grads cancel each other in all-reduce.

            

Reported by Pylint.

                  n = 16
    training_examples = FeatureSet(
        dense_features=torch.zeros((n, D_DENSE)),
        sparse_features=torch.zeros(n, dtype=torch.long),
        values=torch.zeros(n),
    )
    idx = 0
    # Every example has another one that has exactly the same features but an
    # opposite value. Therefore, their grads cancel each other in all-reduce.

            

Reported by Pylint.

                  training_examples = FeatureSet(
        dense_features=torch.zeros((n, D_DENSE)),
        sparse_features=torch.zeros(n, dtype=torch.long),
        values=torch.zeros(n),
    )
    idx = 0
    # Every example has another one that has exactly the same features but an
    # opposite value. Therefore, their grads cancel each other in all-reduce.
    for value in (-1, 1):

            

Reported by Pylint.

                      for x in (-1.0 * value, 1.0 * value):
            for y in (1.0 * value, -1.0 * value):
                for z in (0, 1):
                    training_examples.dense_features[idx, :] = torch.tensor((x, y))
                    training_examples.sparse_features[idx] = z
                    training_examples.values[idx] = value
                    idx += 1

    # Split the examples among NUM_TRAINERS trainers

            

Reported by Pylint.

              from typing import List, Optional

from torch import Tensor
from .module import Module
from .utils import _single, _pair, _triple
from .. import functional as F

from ..common_types import (_size_any_t, _size_1_t, _size_2_t, _size_3_t,
                            _ratio_3_t, _ratio_2_t, _size_any_opt_t, _size_2_opt_t, _size_3_opt_t)

            

Reported by Pylint.

              
from torch import Tensor
from .module import Module
from .utils import _single, _pair, _triple
from .. import functional as F

from ..common_types import (_size_any_t, _size_1_t, _size_2_t, _size_3_t,
                            _ratio_3_t, _ratio_2_t, _size_any_opt_t, _size_2_opt_t, _size_3_opt_t)


            

Reported by Pylint.

              from torch import Tensor
from .module import Module
from .utils import _single, _pair, _triple
from .. import functional as F

from ..common_types import (_size_any_t, _size_1_t, _size_2_t, _size_3_t,
                            _ratio_3_t, _ratio_2_t, _size_any_opt_t, _size_2_opt_t, _size_3_opt_t)



            

Reported by Pylint.

              from .utils import _single, _pair, _triple
from .. import functional as F

from ..common_types import (_size_any_t, _size_1_t, _size_2_t, _size_3_t,
                            _ratio_3_t, _ratio_2_t, _size_any_opt_t, _size_2_opt_t, _size_3_opt_t)


class _MaxPoolNd(Module):
    __constants__ = ['kernel_size', 'stride', 'padding', 'dilation',

            

Reported by Pylint.

                  padding: _size_1_t
    dilation: _size_1_t

    def forward(self, input: Tensor) -> Tensor:
        return F.max_pool1d(input, self.kernel_size, self.stride,
                            self.padding, self.dilation, self.ceil_mode,
                            self.return_indices)



            

Reported by Pylint.

                  padding: _size_2_t
    dilation: _size_2_t

    def forward(self, input: Tensor) -> Tensor:
        return F.max_pool2d(input, self.kernel_size, self.stride,
                            self.padding, self.dilation, self.ceil_mode,
                            self.return_indices)



            

Reported by Pylint.

                  padding: _size_3_t
    dilation: _size_3_t

    def forward(self, input: Tensor) -> Tensor:
        return F.max_pool3d(input, self.kernel_size, self.stride,
                            self.padding, self.dilation, self.ceil_mode,
                            self.return_indices)



            

Reported by Pylint.

                      self.stride = _single(stride if (stride is not None) else kernel_size)
        self.padding = _single(padding)

    def forward(self, input: Tensor, indices: Tensor, output_size: Optional[List[int]] = None) -> Tensor:
        return F.max_unpool1d(input, indices, self.kernel_size, self.stride,
                              self.padding, output_size)


class MaxUnpool2d(_MaxUnpoolNd):

            

Reported by Pylint.

                      self.stride = _pair(stride if (stride is not None) else kernel_size)
        self.padding = _pair(padding)

    def forward(self, input: Tensor, indices: Tensor, output_size: Optional[List[int]] = None) -> Tensor:
        return F.max_unpool2d(input, indices, self.kernel_size, self.stride,
                              self.padding, output_size)


class MaxUnpool3d(_MaxUnpoolNd):

            

Reported by Pylint.

                      self.stride = _triple(stride if (stride is not None) else kernel_size)
        self.padding = _triple(padding)

    def forward(self, input: Tensor, indices: Tensor, output_size: Optional[List[int]] = None) -> Tensor:
        return F.max_unpool3d(input, indices, self.kernel_size, self.stride,
                              self.padding, output_size)


class _AvgPoolNd(Module):

            

Reported by Pylint.

              
import unittest

from hypothesis import given, assume, settings
import hypothesis.strategies as st
import numpy as np
import operator

from caffe2.proto import caffe2_pb2

            

Reported by Pylint.

              import unittest

from hypothesis import given, assume, settings
import hypothesis.strategies as st
import numpy as np
import operator

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

            

Reported by Pylint.

                      op = core.CreateOperator(
            "SumReduceLike", ["X", "Y"], "out", broadcast=1, axis=0)

        def ref_op(X, Y):
            res = np.sum(X, axis=3)
            res = np.sum(res, axis=2)
            return [res]

        self.assertReferenceChecks(

            

Reported by Pylint.

                      op = core.CreateOperator(
            "SumReduceLike", ["X", "Y"], "out", broadcast=1, axis=1)

        def ref_op(X, Y):
            res = np.sum(X, axis=0)
            res = np.sum(res, axis=2)
            return [res]

        self.assertReferenceChecks(

            

Reported by Pylint.

                      op = core.CreateOperator(
            "SumReduceLike", ["X", "Y"], "out", broadcast=1)

        def ref_op(X, Y):
            res = np.sum(X, axis=0)
            res = np.sum(res, axis=2)
            return [res.reshape(Y.shape)]

        self.assertReferenceChecks(

            

Reported by Pylint.

              import caffe2.python.serialized_test.serialized_test_util as serial


# TODO(jiayq): make them hypothesis tests for better coverage.
class TestElementwiseBroadcast(serial.SerializedTestCase):

    def __generate_test_cases(self, allow_broadcast_fastpath: bool):
        """
        generates a set of test cases

            

Reported by Pylint.

              
    @given(**hu.gcs)
    @settings(deadline=None)
    def test_broadcast_powt(self, gc, dc):
        np.random.seed(101)

        #operator
        def powt_op(X, Y):
            return [np.power(X, Y)]

            

Reported by Pylint.

                                                 grad_reference=powt_grad_mixed)

    @given(allow_broadcast_fastpath=st.booleans(), **hu.gcs)
    def test_broadcast_scalar(self, allow_broadcast_fastpath: bool, gc, dc):
        # broadcasting constant
        X = np.random.rand(2, 3, 4, 5).astype(np.float32)
        Y = np.random.rand(1).astype(np.float32)
        op = core.CreateOperator(
            "Add", ["X", "Y"], "out", broadcast=1, allow_broadcast_fastpath=allow_broadcast_fastpath

            

Reported by Pylint.

                      self.assertDeviceChecks(dc, op, [X, Y], [0])

    @given(allow_broadcast_fastpath=st.booleans(), **hu.gcs)
    def test_semantic_broadcast(self, allow_broadcast_fastpath: bool, gc, dc):
        # NCHW as default
        X = np.random.rand(2, 3, 4, 5).astype(np.float32)
        Y = np.random.rand(3).astype(np.float32)
        op = core.CreateOperator(
            "Add", ["X", "Y"], "out", broadcast=1, axis_str="C",

            

Reported by Pylint.

                      self.assertDeviceChecks(dc, op, [X, Y], [0])

    @given(**hu.gcs)
    def test_sum_reduce_empty_blob(self, gc, dc):
        net = core.Net('test')

        with core.DeviceScope(gc):
            net.GivenTensorFill([], ["X"], values=[], shape=[2, 0, 5])
            net.GivenTensorFill([], ["Y"], values=[], shape=[2, 0])

            

Reported by Pylint.