| /external/pytorch/torch/_numpy/ |
| D | _casting_dicts.py | 3 import torch
11 torch.float16: {
12 torch.float16: True,
13 torch.float32: False,
14 torch.float64: False,
15 torch.complex64: False,
16 torch.complex128: False,
17 torch.uint8: False,
18 torch.uint16: False,
19 torch.uint32: False,
[all …]
|
| /external/pytorch/torch/_dynamo/ |
| D | trace_rules.py | 38 import torch
39 import torch._inductor.test_operators
40 import torch.distributed
41 import torch.utils._content_store
42 from torch.utils import _config_module
93 * PyTorch(torch) is in the BUILTIN_SKIPLIST by default, but there are many cases
94 where we want inline the functions under torch namespace.
118 - torch.add: should be put into the FX graph.
119 - torch.is_floating_point: constant folded.
123 For developers: If you add/remove a torch level API, it may trigger failures from
[all …]
|
| /external/pytorch/functorch/dim/ |
| D | op_properties.py | 6 import torch
62 *(getattr(torch.Tensor, m) for m in pointwise_methods),
63 torch.nn.functional.dropout,
64 torch.where,
65 torch.Tensor.abs,
66 torch.abs,
67 torch.Tensor.acos,
68 torch.acos,
69 torch.Tensor.acosh,
70 torch.acosh,
[all …]
|
| /external/pytorch/test/mobile/model_test/ |
| D | math_ops.py | 1 # https://pytorch.org/docs/stable/torch.html#math-operations
5 import torch
8 class PointwiseOpsModule(torch.nn.Module):
13 a = torch.randn(4)
14 b = torch.randn(4)
15 t = torch.tensor([-1, -2, 3], dtype=torch.int8)
16 r = torch.tensor([0, 1, 10, 0], dtype=torch.int8)
17 t = torch.tensor([-1, -2, 3], dtype=torch.int8)
18 s = torch.tensor([4, 0, 1, 0], dtype=torch.int8)
19 f = torch.zeros(3)
[all …]
|
| D | tensor_ops.py | 1 import torch
4 class TensorOpsModule(torch.nn.Module):
9 a = torch.randn(4)
10 b = torch.tensor([1.5])
11 x = torch.ones((2,))
12 c = torch.randn(4, dtype=torch.cfloat)
13 w = torch.rand(4, 4, 4, 4)
14 v = torch.rand(4, 4, 4, 4)
16 # torch.is_tensor(a),
17 # torch.is_storage(a),
[all …]
|
| /external/pytorch/test/typing/pass/ |
| D | math_ops.py | 4 import torch
7 a = torch.randn(4)
8 b = torch.randn(4)
9 t = torch.tensor([-1, -2, 3], dtype=torch.int8)
12 torch.abs(torch.tensor([-1, -2, 3]))
13 torch.absolute(torch.tensor([-1, -2, 3]))
16 torch.acos(a)
17 torch.arccos(a)
20 torch.acosh(a.uniform_(1, 2))
23 torch.add(a, 20)
[all …]
|
| D | creation_ops.py | 5 import torch
6 from torch.testing._internal.common_utils import TEST_NUMPY
15 # torch.tensor()
16 torch.tensor([[0.1, 1.2], [2.2, 3.1], [4.9, 5.2]])
17 torch.tensor([0, 1])
18 torch.tensor(
19 [[0.11111, 0.222222, 0.3333333]], dtype=torch.float64, device=torch.device("cuda:0")
21 torch.tensor(3.14159)
23 # torch.sparse_coo_tensor
24 i = torch.tensor([[0, 1, 1], [2, 0, 2]])
[all …]
|
| /external/pytorch/test/cpp/lazy/ |
| D | test_lazy_ops.cpp | 5 #include <torch/csrc/lazy/core/debug_util.h>
6 #include <torch/csrc/lazy/core/helpers.h>
7 #include <torch/csrc/lazy/core/ir_builder.h>
8 #include <torch/csrc/lazy/core/lazy_graph_executor.h>
9 #include <torch/csrc/lazy/core/metrics.h>
10 #include <torch/csrc/lazy/core/permutation_util.h>
11 #include <torch/csrc/lazy/ts_backend/dynamic_ir.h>
12 #include <torch/csrc/lazy/ts_backend/ts_backend_impl.h>
13 #include <torch/torch.h>
16 namespace torch { namespace
[all …]
|
| /external/pytorch/test/cpp/api/ |
| D | optim_baseline.h | 3 #include <torch/types.h>
9 inline std::vector<std::vector<torch::Tensor>> LBFGS() { in LBFGS()
12 torch::tensor( in LBFGS()
19 torch::tensor( in LBFGS()
21 torch::tensor( in LBFGS()
23 torch::tensor({-0.43108206822505857}), in LBFGS()
26 torch::tensor( in LBFGS()
33 torch::tensor( in LBFGS()
35 torch::tensor( in LBFGS()
37 torch::tensor({-4.776742087865583}), in LBFGS()
[all …]
|
| /external/executorch/backends/qualcomm/quantizer/ |
| D | qconfig.py | 4 import torch
5 from torch import Tensor
6 from torch.ao.quantization.fake_quantize import (
10 from torch.ao.quantization.observer import (
16 from torch.ao.quantization.quantizer import DerivedQuantizationSpec, QuantizationSpec
17 from torch.fx import Node
39 (broadcast_act_scale, broadcast_weight_scale) = torch.broadcast_tensors(
42 derived_scale = (broadcast_act_scale * broadcast_weight_scale).to(torch.float32)
43 derived_zero = torch.zeros(derived_scale.size()).to(torch.int32)
54 dtype=torch.int32,
[all …]
|
| /external/pytorch/test/inductor/ |
| D | test_b2b_gemm.py | 5 import torch
6 from torch._inductor.runtime.benchmarking import benchmarker
7 from torch._inductor.test_case import run_tests, TestCase
8 from torch._inductor.utils import run_and_get_code
9 from torch.testing._internal.inductor_utils import HAS_CUDA
13 @torch._dynamo.config.patch(cache_size_limit=32)
14 @torch._inductor.config.patch(b2b_gemm_pass=True)
21 def f(m1: torch.Tensor, m2: torch.Tensor, m3: torch.Tensor) -> torch.Tensor:
22 g = torch.nn.GELU()
23 return torch.mm(g(torch.mm(m1, m2)), m3)
[all …]
|
| /external/executorch/backends/arm/test/ops/ |
| D | test_batch_norm.py | 13 import torch
25 torch.zeros(1, 32, 112, 112),
34 torch.zeros(1, 32, 112, 112),
43 torch.zeros(1, 32, 112, 112),
48 torch.rand(32),
49 torch.rand(32),
50 torch.rand(32),
51 torch.rand(32),
56 torch.zeros(1, 32, 112, 112),
61 torch.rand(32),
[all …]
|
| /external/pytorch/test/ |
| D | test_type_promotion.py | 7 import torch
9 from torch.testing._internal.common_utils import (TestCase, run_tests, load_tests, make_tensor,
13 from torch.testing._internal.common_device_type import (instantiate_device_type_tests, onlyNativeDe…
15 from torch.testing._internal.common_dtype import (
24 # load_tests from torch.testing._internal.common_utils is used to automatically filter tests for
29 # the default dtype being torch.float and again with the default dtype
30 # being torch.double.
34 with set_default_dtype(torch.float):
36 with set_default_dtype(torch.double):
51 int_tensor = torch.ones([4, 4, 4], dtype=torch.int32, device=device)
[all …]
|
| /external/pytorch/torch/csrc/jit/runtime/ |
| D | serialized_shape_function_registry.cpp | 9 #include <torch/csrc/jit/jit_log.h>
10 #include <torch/csrc/jit/passes/inliner.h>
11 #include <torch/csrc/jit/runtime/operator.h>
12 #include <torch/csrc/jit/runtime/serialized_shape_function_registry.h>
16 namespace torch::jit { namespace
23 for _0 in range(torch.len(self)):
25 _1 = torch.append(out, elem)
30 if torch.eq(torch.len(out), 2):
34 if torch.eq(torch.len(self), 3):
37 _0 = torch.eq(torch.len(self), 4)
[all …]
|
| /external/pytorch/test/dynamo/ |
| D | test_ctx_manager.py | 4 import torch
5 import torch._dynamo.test_case
6 import torch._dynamo.testing
7 import torch.onnx.operators
8 from torch._dynamo.testing import EagerAndRecordGraphs, normalize_gm, same
9 from torch.nn import functional as F
10 from torch.testing._internal.common_cuda import PLATFORM_SUPPORTS_FLASH_ATTENTION
11 from torch.testing._internal.common_utils import TEST_WITH_ROCM
16 self.prev = torch.is_grad_enabled()
20 torch._C._set_grad_enabled(self.mode)
[all …]
|
| D | test_trace_rules.py | 11 import torch
12 import torch._dynamo.config as config
13 import torch._dynamo.test_case
14 import torch._functorch.deprecated as deprecated_func
15 from torch._dynamo.trace_rules import (
23 from torch._dynamo.utils import hashable, is_safe_constant, istype
24 from torch._dynamo.variables import TorchInGraphFunctionVariable, UserFunctionVariable
25 from torch.testing._internal.common_utils import skipIfWindows
36 "torch._nested_tensor_from_mask",
37 "torch._nested_from_padded",
[all …]
|
| /external/pytorch/torch/ao/quantization/pt2e/representation/ |
| D | rewrite.py | 6 import torch
7 from torch._higher_order_ops.out_dtype import out_dtype
8 from torch.ao.quantization.fx._decomposed import quantized_decomposed_lib # noqa: F401
9 from torch.ao.quantization.pt2e.export_utils import _WrapperModule
10 from torch.ao.quantization.pt2e.utils import (
16 from torch.fx import GraphModule
17 from torch.fx.subgraph_rewriter import replace_pattern
26 torch.randint(-128, 127, (2, 5), dtype=torch.int8),
27 torch.randn(1, dtype=torch.float),
28 torch.zeros(1, dtype=torch.int),
[all …]
|
| /external/pytorch/test/jit/ |
| D | test_batch_mm.py | 3 import torch
4 from torch.testing import FileCheck
5 from torch.testing._internal.jit_utils import JitTestCase
20 torch.tensor([[1 + x, 2 + x, 3 + x], [4 + x, 5 + x, 6 + x]])
22 else torch.tensor([[1 + x, 2 + x], [3 + x, 4 + x], [5 + x, 6 + x]])
28 T1: torch.Tensor,
29 T2: torch.Tensor,
30 T3: torch.Tensor,
31 T4: torch.Tensor,
32 T5: torch.Tensor,
[all …]
|
| /external/pytorch/ |
| D | build_variables.bzl | 21 "torch/csrc/autograd/generated/Functions.cpp",
22 "torch/csrc/autograd/generated/VariableType_0.cpp",
23 "torch/csrc/autograd/generated/VariableType_1.cpp",
24 "torch/csrc/autograd/generated/VariableType_2.cpp",
25 "torch/csrc/autograd/generated/VariableType_3.cpp",
26 "torch/csrc/autograd/generated/VariableType_4.cpp",
27 "torch/csrc/autograd/generated/ViewFuncs.cpp",
28 "torch/csrc/autograd/generated/TraceType_0.cpp",
29 "torch/csrc/autograd/generated/TraceType_1.cpp",
30 "torch/csrc/autograd/generated/TraceType_2.cpp",
[all …]
|
| /external/pytorch/torch/_prims/ |
| D | context.py | 6 import torch
7 import torch._decomp
8 import torch._prims
9 import torch._refs
10 import torch._refs.nn
11 import torch._refs.nn.functional
12 import torch._refs.special
13 import torch.overrides
14 from torch._prims_common import torch_function_passthrough
20 Mapping of torch API functions to torch._refs functions.
[all …]
|
| /external/pytorch/test/cpp_api_parity/ |
| D | parity-tracker.md | 3 ## torch::nn
6 torch::nn::Sequential|Yes|No
7 torch::nn::ModuleList|Yes|No
8 torch::nn::ModuleDict|No|No
9 torch::nn::ParameterList|No|No
10 torch::nn::ParameterDict|No|No
11 torch::nn::Conv1d|Yes|No
12 torch::nn::Conv2d|Yes|No
13 torch::nn::Conv3d|Yes|No
14 torch::nn::ConvTranspose1d|Yes|No
[all …]
|
| /external/executorch/kernels/quantized/test/ |
| D | test_quant_dequant_per_token.py | 11 import torch
12 from torch.ao.quantization.fx._decomposed import quantized_decomposed_lib # noqa: F401
18 input_tensor = torch.tensor(
19 [[-0.5, 0.3, 1.2], [0.1, -0.8, 2.1], [-5, 1, 2]], dtype=torch.float32
21 scale = torch.tensor([0.5, 0.8, 1.0], dtype=torch.float64)
23 zero_point = torch.tensor([-1, -2, 0])
25 quantized_tensor = torch.ops.quantized_decomposed.quantize_per_token(
26 input_tensor, scale, zero_point, -128, 127, torch.int8
28 expected_quantized_tensor = torch.ops.et_quant_test.quantize_per_token(
29 input_tensor, scale, zero_point, -128, 127, torch.int8
[all …]
|
| /external/pytorch/torch/ |
| D | overrides.py | 4 While most of the torch API and handling for ``__torch_function__`` happens
5 at the C++ level, some of the torch API is written in Python so we need
8 has_torch_function. See torch/functional.py and test/test_overrides.py
33 import torch
34 from torch._C import (
64 module: str = "torch",
105 A tuple of functions that are publicly available in the torch API but cannot
111 >>> torch.Tensor.as_subclass in torch.overrides.get_ignored_functions()
113 >>> torch.add in torch.overrides.get_ignored_functions()
116 Tensor = torch.Tensor
[all …]
|
| /external/pytorch/torch/testing/_internal/ |
| D | autocast_test_lists.py | 5 import torch
6 from torch.testing._internal.common_utils import TEST_WITH_ROCM
7 from torch.testing._internal.common_utils import TestCase
12 input = (torch.randn((n, n), device=dev, dtype=torch.float32),)
14 hx = ((torch.randn((n, n), device=dev, dtype=torch.float32),
15 torch.randn((n, n), device=dev, dtype=torch.float32)) if is_lstm else
16 torch.randn((n, n), device=dev, dtype=torch.float32),)
18 weights = (torch.randn((num_chunks * n, n), device=dev, dtype=torch.float32), # weight_ih
19 torch.randn((num_chunks * n, n), device=dev, dtype=torch.float32), # weight_hh
20 torch.randn((num_chunks * n), device=dev, dtype=torch.float32), # bias_ih
[all …]
|
| /external/pytorch/torch/ao/nn/quantized/reference/modules/ |
| D | utils.py | 4 import torch
12 class ReferenceQuantizedModule(torch.nn.Module):
16 "qscheme": torch.per_tensor_affine,
17 "dtype": torch.quint8,
21 self.weight_qscheme: torch.qscheme = weight_qparams["qscheme"]
25 torch.per_tensor_affine,
26 torch.per_channel_affine,
27 torch.per_channel_affine_float_qparams,
30 torch.quint8,
31 torch.qint8,
[all …]
|