import os
import sys

import torch
from torch import nn

# 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.

              import sys

import torch
from torch import nn

# 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)
from torch.testing._internal.jit_utils import JitTestCase

            

Reported by Pylint.

              # 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)
from torch.testing._internal.jit_utils import JitTestCase

if __name__ == '__main__':
    raise RuntimeError("This test file is not meant to be run directly, use:\n\n"
                       "\tpython test/test_jit.py TESTNAME\n\n"
                       "instead.")

            

Reported by Pylint.

                      self.lstm2 = nn.LSTMCell(51, 51)
        self.linear = nn.Linear(51, 1)

    def forward(self, input):
        outputs = []
        h_t = torch.zeros(input.size(0), 51)
        c_t = torch.zeros(input.size(0), 51)
        h_t2 = torch.zeros(input.size(0), 51)
        c_t2 = torch.zeros(input.size(0), 51)

            

Reported by Pylint.

              
    def test_basic(self):
        seq = torch.jit.script(Sequence())
        p = torch.jit._ScriptProfile()
        p.enable()
        seq(torch.rand((10, 100)))
        p.disable()
        self.assertNotEqual(p.dump_string(), "")


            

Reported by Pylint.

              
        @torch.jit.script
        def fn():
            p = torch.jit._ScriptProfile()
            p.enable()
            _ = seq(torch.rand((10, 100)))
            p.disable()
            return p


            

Reported by Pylint.

              
    def test_multi(self):
        seq = torch.jit.script(Sequence())
        profiles = [torch.jit._ScriptProfile() for _ in range(5)]
        for p in profiles:
            p.enable()

        last = None
        while len(profiles) > 0:

            

Reported by Pylint.

              
        @torch.jit.script
        def fn():
            p = torch.jit._ScriptProfile()
            p.enable()
            _ = seq(torch.rand((10, 100)))
            p.disable()
            stats0 = p.dump_string()


            

Reported by Pylint.

                      self.assertNotEqual(s1, s2)

    def test_empty(self):
        p = torch.jit._ScriptProfile()
        p.enable()
        p.disable()
        self.assertEqual(p.dump_string(), "")

            

Reported by Pylint.

              import os
import sys

import torch
from torch import nn

# 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.

              import numpy as np
import caffe2.python.fakelowp.init_shared_libs  # noqa
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial
import datetime
from hypothesis import settings


            

Reported by Pylint.

              import numpy as np
import caffe2.python.fakelowp.init_shared_libs  # noqa
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial
import datetime
from hypothesis import settings


            

Reported by Pylint.

              import numpy as np
import caffe2.python.fakelowp.init_shared_libs  # noqa
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial
import datetime
from hypothesis import settings


            

Reported by Pylint.

              import caffe2.python.fakelowp.init_shared_libs  # noqa
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial
import datetime
from hypothesis import settings

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

            

Reported by Pylint.

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

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


            

Reported by Pylint.

              from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial
import datetime
from hypothesis import settings

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

class DeqSwishQuantTest(serial.SerializedTestCase):
    def _get_scale_zp(self, tensor):

            

Reported by Pylint.

              import numpy as np
import caffe2.python.fakelowp.init_shared_libs  # noqa
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial
import datetime
from hypothesis import settings


            

Reported by Pylint.

                          1 if o.type == "Onnxifi" else 0 for o in net_onnxified.op
        )
        np.testing.assert_equal(num_onnxified_ops, 1)
        # TODO: add an assertion to check the optimized net
        # fused Dequantize->Swish->Quantize to QuantizedSwish
        workspace.CreateNet(net_onnxified)
        workspace.RunNet(net_onnxified.name)
        Y_glow = workspace.FetchInt8Blob("Y")
        U_int8 = workspace.FetchInt8Blob("U_int8")

            

Reported by Pylint.

              import numpy as np
import caffe2.python.fakelowp.init_shared_libs  # noqa
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial
import datetime
from hypothesis import settings


            

Reported by Pylint.

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

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

class DeqSwishQuantTest(serial.SerializedTestCase):

            

Reported by Pylint.

              # Must happen before importing caffe2.python.*
import caffe2.python.fakelowp.init_shared_libs  # noqa
import datetime
import numpy as np
from hypothesis import given, settings, example
from hypothesis import strategies as st
from caffe2.python import core, 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
import datetime
import numpy as np
from hypothesis import given, settings, example
from hypothesis import strategies as st
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial

            

Reported by Pylint.

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


            

Reported by Pylint.

              import numpy as np
from hypothesis import given, settings, example
from hypothesis import strategies as st
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial

# Test that parallel chunks behave the same way as the serial one

            

Reported by Pylint.

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

# Test that parallel chunks behave the same way as the serial one


            

Reported by Pylint.

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

# Test that parallel chunks behave the same way as the serial one

workspace.GlobalInit(

            

Reported by Pylint.

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

# Test that parallel chunks behave the same way as the serial one

workspace.GlobalInit(
    [

            

Reported by Pylint.

              # Must happen before importing caffe2.python.*
import caffe2.python.fakelowp.init_shared_libs  # noqa
import datetime
import numpy as np
from hypothesis import given, settings, example
from hypothesis import strategies as st
from caffe2.python import core, 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
import datetime
import numpy as np
from hypothesis import given, settings, example
from hypothesis import strategies as st
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import caffe2.python.serialized_test.serialized_test_util as serial

            

Reported by Pylint.

                  )
    # @example(m=64, k=5423, n=553, scale=1e-3, zp=120, rand_seed=1)
    @settings(deadline=datetime.timedelta(seconds=1000), max_examples=1)
    def test_ParallelFC(self, m, k, n, scale, zp, rand_seed):
        np.random.seed(rand_seed)
        workspace.ResetWorkspace()

        # Y = W_T * X + b
        X_fp32 = np.random.uniform(-1, 1, size=(m, k)).astype(np.float16) \

            

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 caffe2.python.serialized_test.serialized_test_util as serial

            

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 caffe2.python.serialized_test.serialized_test_util as serial
import datetime

            

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 caffe2.python.serialized_test.serialized_test_util as serial
import datetime


            

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 caffe2.python.serialized_test.serialized_test_util as serial
import datetime

core.GlobalInit(["caffe2", "--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 caffe2.python.serialized_test.serialized_test_util as serial
import datetime

core.GlobalInit(["caffe2", "--glow_global_fp16=1",
                 "--glow_global_fused_scale_offset_fp16=1",

            

Reported by Pylint.

              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 caffe2.python.serialized_test.serialized_test_util as serial
import datetime

core.GlobalInit(["caffe2", "--glow_global_fp16=1",
                 "--glow_global_fused_scale_offset_fp16=1",
                 "--glow_global_force_sls_fp16_accum=1"])

            

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 numpy as np
import unittest
import caffe2.python.fakelowp.init_shared_libs  # noqa

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

              import unittest
import caffe2.python.fakelowp.init_shared_libs  # noqa

from caffe2.proto import caffe2_pb2
from caffe2.python import core, workspace
from caffe2.python.onnx.onnxifi import onnxifi_caffe2_net
from caffe2.python.fakelowp.test_utils import print_test_debug_info
import datetime
from hypothesis import given, settings

            

Reported by Pylint.

              import caffe2.python.fakelowp.init_shared_libs  # noqa

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

            

Reported by Pylint.

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

            

Reported by Pylint.

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


            

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
from hypothesis import given, settings
import hypothesis.strategies as st
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.fakelowp.test_utils import print_test_debug_info
import datetime
from hypothesis import given, settings
import hypothesis.strategies as st
import caffe2.python.serialized_test.serialized_test_util as serial

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



            

Reported by Pylint.

              import datetime
from hypothesis import given, settings
import hypothesis.strategies as st
import caffe2.python.serialized_test.serialized_test_util as serial

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


class TestBatchMatMul(serial.SerializedTestCase):

            

Reported by Pylint.

              import numpy as np
import unittest
import caffe2.python.fakelowp.init_shared_libs  # noqa

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

              import numpy as np
import unittest
import caffe2.python.fakelowp.init_shared_libs  # noqa

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

              import re


"""
Example Usages

Create a new commitlist for consumption by categorize.py.
Said commitlist contains commits between v1.5.0 and f5bc91f851.


            

Reported by Pylint.

              import argparse
from common import run, topics
from collections import defaultdict
import os
import csv
import pprint
from common import CommitDataCache
import re


            

Reported by Pylint.

              import argparse
from common import run, topics
from collections import defaultdict
import os
import csv
import pprint
from common import CommitDataCache
import re


            

Reported by Pylint.

              import argparse
from common import run, topics
from collections import defaultdict
import os
import csv
import pprint
from common import CommitDataCache
import re


            

Reported by Pylint.

              from common import run, topics
from collections import defaultdict
import os
import csv
import pprint
from common import CommitDataCache
import re



            

Reported by Pylint.

              from collections import defaultdict
import os
import csv
import pprint
from common import CommitDataCache
import re


"""

            

Reported by Pylint.

              import csv
import pprint
from common import CommitDataCache
import re


"""
Example Usages


            

Reported by Pylint.

              
"""

class Commit:
    def __init__(self, commit_hash, category, topic, title):
        self.commit_hash = commit_hash
        self.category = category
        self.topic = topic
        self.title = title

            

Reported by Pylint.

                  def __repr__(self):
        return f'Commit({self.commit_hash}, {self.category}, {self.topic}, {self.title})'

class CommitList:
    # NB: Private ctor. Use `from_existing` or `create_new`.
    def __init__(self, path, commits):
        self.path = path
        self.commits = commits


            

Reported by Pylint.

                      self.commits = commits

    @staticmethod
    def from_existing(path):
        commits = CommitList.read_from_disk(path)
        return CommitList(path, commits)

    @staticmethod
    def create_new(path, base_version, new_version):

            

Reported by Pylint.

                      self.dim = dim
        self.concentration, = broadcast_all(concentration)
        batch_shape = self.concentration.size()
        event_shape = torch.Size((dim, dim))
        # This is used to draw vectorized samples from the beta distribution in Sec. 3.2 of [1].
        marginal_conc = self.concentration + 0.5 * (self.dim - 2)
        offset = torch.arange(self.dim - 1, dtype=self.concentration.dtype, device=self.concentration.device)
        offset = torch.cat([offset.new_zeros((1,)), offset])
        beta_conc1 = offset + 0.5

            

Reported by Pylint.

                      event_shape = torch.Size((dim, dim))
        # This is used to draw vectorized samples from the beta distribution in Sec. 3.2 of [1].
        marginal_conc = self.concentration + 0.5 * (self.dim - 2)
        offset = torch.arange(self.dim - 1, dtype=self.concentration.dtype, device=self.concentration.device)
        offset = torch.cat([offset.new_zeros((1,)), offset])
        beta_conc1 = offset + 0.5
        beta_conc0 = marginal_conc.unsqueeze(-1) - 0.5 * offset
        self._beta = Beta(beta_conc1, beta_conc0)
        super(LKJCholesky, self).__init__(batch_shape, event_shape, validate_args)

            

Reported by Pylint.

                      # This is used to draw vectorized samples from the beta distribution in Sec. 3.2 of [1].
        marginal_conc = self.concentration + 0.5 * (self.dim - 2)
        offset = torch.arange(self.dim - 1, dtype=self.concentration.dtype, device=self.concentration.device)
        offset = torch.cat([offset.new_zeros((1,)), offset])
        beta_conc1 = offset + 0.5
        beta_conc0 = marginal_conc.unsqueeze(-1) - 0.5 * offset
        self._beta = Beta(beta_conc1, beta_conc0)
        super(LKJCholesky, self).__init__(batch_shape, event_shape, validate_args)


            

Reported by Pylint.

              
    def expand(self, batch_shape, _instance=None):
        new = self._get_checked_instance(LKJCholesky, _instance)
        batch_shape = torch.Size(batch_shape)
        new.dim = self.dim
        new.concentration = self.concentration.expand(batch_shape)
        new._beta = self._beta.expand(batch_shape + (self.dim,))
        super(LKJCholesky, new).__init__(batch_shape, self.event_shape, validate_args=False)
        new._validate_args = self._validate_args

            

Reported by Pylint.

                      new._validate_args = self._validate_args
        return new

    def sample(self, sample_shape=torch.Size()):
        # This uses the Onion method, but there are a few differences from [1] Sec. 3.2:
        # - This vectorizes the for loop and also works for heterogeneous eta.
        # - Same algorithm generalizes to n=1.
        # - The procedure is simplified since we are sampling the cholesky factor of
        #   the correlation matrix instead of the correlation matrix itself. As such,

            

Reported by Pylint.

                      #   the correlation matrix instead of the correlation matrix itself. As such,
        #   we only need to generate `w`.
        y = self._beta.sample(sample_shape).unsqueeze(-1)
        u_normal = torch.randn(self._extended_shape(sample_shape),
                               dtype=y.dtype,
                               device=y.device).tril(-1)
        u_hypersphere = u_normal / u_normal.norm(dim=-1, keepdim=True)
        # Replace NaNs in first row
        u_hypersphere[..., 0, :].fill_(0.)

            

Reported by Pylint.

                      u_hypersphere = u_normal / u_normal.norm(dim=-1, keepdim=True)
        # Replace NaNs in first row
        u_hypersphere[..., 0, :].fill_(0.)
        w = torch.sqrt(y) * u_hypersphere
        # Fill diagonal elements; clamp for numerical stability
        eps = torch.finfo(w.dtype).tiny
        diag_elems = torch.clamp(1 - torch.sum(w**2, dim=-1), min=eps).sqrt()
        w += torch.diag_embed(diag_elems)
        return w

            

Reported by Pylint.

                      u_hypersphere[..., 0, :].fill_(0.)
        w = torch.sqrt(y) * u_hypersphere
        # Fill diagonal elements; clamp for numerical stability
        eps = torch.finfo(w.dtype).tiny
        diag_elems = torch.clamp(1 - torch.sum(w**2, dim=-1), min=eps).sqrt()
        w += torch.diag_embed(diag_elems)
        return w

    def log_prob(self, value):

            

Reported by Pylint.

                      w = torch.sqrt(y) * u_hypersphere
        # Fill diagonal elements; clamp for numerical stability
        eps = torch.finfo(w.dtype).tiny
        diag_elems = torch.clamp(1 - torch.sum(w**2, dim=-1), min=eps).sqrt()
        w += torch.diag_embed(diag_elems)
        return w

    def log_prob(self, value):
        # See: https://mc-stan.org/docs/2_25/functions-reference/cholesky-lkj-correlation-distribution.html

            

Reported by Pylint.

                      w = torch.sqrt(y) * u_hypersphere
        # Fill diagonal elements; clamp for numerical stability
        eps = torch.finfo(w.dtype).tiny
        diag_elems = torch.clamp(1 - torch.sum(w**2, dim=-1), min=eps).sqrt()
        w += torch.diag_embed(diag_elems)
        return w

    def log_prob(self, value):
        # See: https://mc-stan.org/docs/2_25/functions-reference/cholesky-lkj-correlation-distribution.html

            

Reported by Pylint.

              import collections

import caffe2.python.hypothesis_test_util as hu
import hypothesis.strategies as st
import numpy as np
from caffe2.python import core, dyndep, workspace
from caffe2.quantization.server.dnnlowp_test_utils import check_quantized_results_close
from hypothesis import given, settings


            

Reported by Pylint.

              import numpy as np
from caffe2.python import core, dyndep, workspace
from caffe2.quantization.server.dnnlowp_test_utils import check_quantized_results_close
from hypothesis import given, settings


dyndep.InitOpsLibrary("//caffe2/caffe2/quantization/server:dnnlowp_ops")
workspace.GlobalInit(["caffe2", "--caffe2_omp_num_threads=11"])


            

Reported by Pylint.

                  )
    @settings(deadline=None)
    def test_dnnlowp_elementwise_mul_int(
        self, N, is_empty, in_quantized, out_quantized, in_place, gc, dc
    ):
        if is_empty:
            N = 0
        # FIXME: DNNLOWP Mul doesn't support inplace operation and
        # dequantize_output=1 at the same time

            

Reported by Pylint.

                  ):
        if is_empty:
            N = 0
        # FIXME: DNNLOWP Mul doesn't support inplace operation and
        # dequantize_output=1 at the same time
        if in_place[0] or in_place[1]:
            in_quantized = True
            out_quantized = True


            

Reported by Pylint.

              
    @given(**hu.gcs_cpu_only)
    @settings(deadline=None)
    def test_dnnlowp_elementwise_mul_broadcast(self, gc, dc):
        # Set broadcast and no axis, i.e. broadcasting last dimensions.
        min_ = -100
        max_ = min_ + 255
        A = np.round(np.random.rand(2, 3, 4, 5) * (max_ - min_) + min_)
        A = A.astype(np.float32)

            

Reported by Pylint.

              
    @given(**hu.gcs_cpu_only)
    @settings(deadline=None)
    def test_dnnlowp_elementwise_mul_broadcast_axis(self, gc, dc):
        for bdim, axis in [
            ((3, 4), 1),  # broadcasting intermediate dimensions
            ((2,), 0),  # broadcasting the first dimension
            ((1, 4, 1), 1),
        ]:

            

Reported by Pylint.

              

import collections

import caffe2.python.hypothesis_test_util as hu
import hypothesis.strategies as st
import numpy as np
from caffe2.python import core, dyndep, workspace
from caffe2.quantization.server.dnnlowp_test_utils import check_quantized_results_close

            

Reported by Pylint.

              workspace.GlobalInit(["caffe2", "--caffe2_omp_num_threads=11"])


class DNNLowPMulOpTest(hu.HypothesisTestCase):
    @given(
        N=st.integers(32, 256),
        is_empty=st.booleans(),
        in_quantized=st.booleans(),
        out_quantized=st.booleans(),

            

Reported by Pylint.

                      in_place=st.sampled_from([(False, False), (True, False), (False, True)]),
        **hu.gcs_cpu_only
    )
    @settings(deadline=None)
    def test_dnnlowp_elementwise_mul_int(
        self, N, is_empty, in_quantized, out_quantized, in_place, gc, dc
    ):
        if is_empty:
            N = 0

            

Reported by Pylint.

                      in_place=st.sampled_from([(False, False), (True, False), (False, True)]),
        **hu.gcs_cpu_only
    )
    @settings(deadline=None)
    def test_dnnlowp_elementwise_mul_int(
        self, N, is_empty, in_quantized, out_quantized, in_place, gc, dc
    ):
        if is_empty:
            N = 0

            

Reported by Pylint.

              import torch.autograd
import torch.cuda

from . import microbatch
from .batchnorm import DeferredBatchNorm
from .pipeline import Pipeline
from .skip.layout import inspect_skip_layout
from .skip.skippable import verify_skippables
from .stream import AbstractStream, new_stream

            

Reported by Pylint.

              import torch.cuda

from . import microbatch
from .batchnorm import DeferredBatchNorm
from .pipeline import Pipeline
from .skip.layout import inspect_skip_layout
from .skip.skippable import verify_skippables
from .stream import AbstractStream, new_stream


            

Reported by Pylint.

              
from . import microbatch
from .batchnorm import DeferredBatchNorm
from .pipeline import Pipeline
from .skip.layout import inspect_skip_layout
from .skip.skippable import verify_skippables
from .stream import AbstractStream, new_stream

__all__ = ["Pipe"]

            

Reported by Pylint.

              from . import microbatch
from .batchnorm import DeferredBatchNorm
from .pipeline import Pipeline
from .skip.layout import inspect_skip_layout
from .skip.skippable import verify_skippables
from .stream import AbstractStream, new_stream

__all__ = ["Pipe"]


            

Reported by Pylint.

              from .batchnorm import DeferredBatchNorm
from .pipeline import Pipeline
from .skip.layout import inspect_skip_layout
from .skip.skippable import verify_skippables
from .stream import AbstractStream, new_stream

__all__ = ["Pipe"]



            

Reported by Pylint.

              from .pipeline import Pipeline
from .skip.layout import inspect_skip_layout
from .skip.skippable import verify_skippables
from .stream import AbstractStream, new_stream

__all__ = ["Pipe"]


Device = Union[torch.device, int, str]

            

Reported by Pylint.

              __all__ = ["Pipe"]


Device = Union[torch.device, int, str]
Devices = Union[Iterable[Device], List[Device]]

Tensors = Sequence[Tensor]
TensorOrTensors = Union[Tensor, Tensors]


            

Reported by Pylint.

              
if TYPE_CHECKING:
    # Typechecking: nn.Module is not a Generic
    Module = nn.Module[TensorOrTensors]  # type: ignore[type-arg]
    NamedModules = OrderedDict[str, Module]
else:
    Module = nn.Module
    NamedModules = OrderedDict


            

Reported by Pylint.

              if TYPE_CHECKING:
    # Typechecking: nn.Module is not a Generic
    Module = nn.Module[TensorOrTensors]  # type: ignore[type-arg]
    NamedModules = OrderedDict[str, Module]
else:
    Module = nn.Module
    NamedModules = OrderedDict



            

Reported by Pylint.

              

def _verify_splitting(
    module: nn.Sequential, partitions: List[nn.Sequential], devices: List[torch.device]
) -> None:
    num_parameters = len(list(module.parameters()))
    num_child_parameters = sum(len(list(child.parameters())) for child in module.children())
    if num_parameters == num_child_parameters:
        return

            

Reported by Pylint.

              def is_sparse(A):
    """Check if tensor A is a sparse tensor"""
    if isinstance(A, torch.Tensor):
        return A.layout == torch.sparse_coo

    error_str = "expected Tensor"
    if not torch.jit.is_scripting():
        error_str += " but got {}".format(type(A))
    raise TypeError(error_str)

            

Reported by Pylint.

                  Integer types map to float32.
    """
    dtype = A.dtype
    if dtype in (torch.float16, torch.float32, torch.float64):
        return dtype
    return torch.float32


def matmul(A: Optional[Tensor], B: Tensor) -> Tensor:

            

Reported by Pylint.

                  Integer types map to float32.
    """
    dtype = A.dtype
    if dtype in (torch.float16, torch.float32, torch.float64):
        return dtype
    return torch.float32


def matmul(A: Optional[Tensor], B: Tensor) -> Tensor:

            

Reported by Pylint.

                  Integer types map to float32.
    """
    dtype = A.dtype
    if dtype in (torch.float16, torch.float32, torch.float64):
        return dtype
    return torch.float32


def matmul(A: Optional[Tensor], B: Tensor) -> Tensor:

            

Reported by Pylint.

                  dtype = A.dtype
    if dtype in (torch.float16, torch.float32, torch.float64):
        return dtype
    return torch.float32


def matmul(A: Optional[Tensor], B: Tensor) -> Tensor:
    """Multiply two matrices.


            

Reported by Pylint.

                      return B
    if is_sparse(A):
        return torch.sparse.mm(A, B)
    return torch.matmul(A, B)


def conjugate(A):
    """Return conjugate of tensor A.


            

Reported by Pylint.

                      # torch.orgqr is not available in CUDA
        Q = torch.linalg.qr(A).Q
    else:
        Q = torch.orgqr(*torch.geqrf(A))
    return Q


def symeig(A: Tensor, largest: Optional[bool] = False) -> Tuple[Tensor, Tensor]:
    """Return eigenpairs of A with specified ordering.

            

Reported by Pylint.

                      # torch.orgqr is not available in CUDA
        Q = torch.linalg.qr(A).Q
    else:
        Q = torch.orgqr(*torch.geqrf(A))
    return Q


def symeig(A: Tensor, largest: Optional[bool] = False) -> Tuple[Tensor, Tensor]:
    """Return eigenpairs of A with specified ordering.

            

Reported by Pylint.

                  E, Z = torch.linalg.eigh(A, UPLO='U')
    # assuming that E is ordered
    if largest:
        E = torch.flip(E, dims=(-1,))
        Z = torch.flip(Z, dims=(-1,))
    return E, Z

            

Reported by Pylint.

                  # assuming that E is ordered
    if largest:
        E = torch.flip(E, dims=(-1,))
        Z = torch.flip(Z, dims=(-1,))
    return E, Z

            

Reported by Pylint.