Home
last modified time | relevance | path

Searched full:mat1 (Results 1 – 25 of 191) sorted by relevance

12345678

/aosp_15_r20/external/eigen/unsupported/test/
H A Dcxx11_tensor_contraction.cpp22 Tensor<float, 2, DataLayout> mat1(2, 3); in test_evals() local
26 mat1.setRandom(); in test_evals()
33 typedef TensorEvaluator<decltype(mat1.contract(mat2, dims3)), DefaultDevice> Evaluator; in test_evals()
34 Evaluator eval(mat1.contract(mat2, dims3), DefaultDevice()); in test_evals()
40 VERIFY_IS_APPROX(mat4(0,0), mat1(0,0)*mat2(0,0) + mat1(1,0)*mat2(1,0)); in test_evals()
41 VERIFY_IS_APPROX(mat4(0,1), mat1(0,0)*mat2(0,1) + mat1(1,0)*mat2(1,1)); in test_evals()
42 VERIFY_IS_APPROX(mat4(0,2), mat1(0,0)*mat2(0,2) + mat1(1,0)*mat2(1,2)); in test_evals()
43 VERIFY_IS_APPROX(mat4(1,0), mat1(0,1)*mat2(0,0) + mat1(1,1)*mat2(1,0)); in test_evals()
44 VERIFY_IS_APPROX(mat4(1,1), mat1(0,1)*mat2(0,1) + mat1(1,1)*mat2(1,1)); in test_evals()
45 VERIFY_IS_APPROX(mat4(1,2), mat1(0,1)*mat2(0,2) + mat1(1,1)*mat2(1,2)); in test_evals()
[all …]
H A Dcxx11_tensor_expr.cpp74 TensorMap<Tensor<float, 2>> mat1(data1, 2, 3); in test_2d() local
78 mat1(0,0) = 0.0; in test_2d()
79 mat1(0,1) = 1.0; in test_2d()
80 mat1(0,2) = 2.0; in test_2d()
81 mat1(1,0) = 3.0; in test_2d()
82 mat1(1,1) = 4.0; in test_2d()
83 mat1(1,2) = 5.0; in test_2d()
94 mat3 = mat1.abs(); in test_2d()
114 Tensor<float, 3> mat1(2,3,7); in test_3d() local
121 mat1(i,j,k) = val; in test_3d()
[all …]
H A Dcxx11_tensor_fixed_size.cpp133 TensorMap<TensorFixedSize<float, Sizes<2, 3> > > mat1(data1,2,3); in test_2d() local
137 VERIFY_IS_EQUAL((mat1.size()), 2*3); in test_2d()
138 VERIFY_IS_EQUAL(mat1.rank(), 2); in test_2d()
139 // VERIFY_IS_EQUAL((mat1.dimension(0)), 2); in test_2d()
140 // VERIFY_IS_EQUAL((mat1.dimension(1)), 3); in test_2d()
142 mat1(0,0) = 0.0; in test_2d()
143 mat1(0,1) = 1.0; in test_2d()
144 mat1(0,2) = 2.0; in test_2d()
145 mat1(1,0) = 3.0; in test_2d()
146 mat1(1,1) = 4.0; in test_2d()
[all …]
H A Dcxx11_tensor_comparisons.cpp19 Tensor<float, 3> mat1(2,3,7); in test_orderings() local
26 mat1.setRandom(); in test_orderings()
29 lt = mat1 < mat2; in test_orderings()
30 le = mat1 <= mat2; in test_orderings()
31 gt = mat1 > mat2; in test_orderings()
32 ge = mat1 >= mat2; in test_orderings()
37 VERIFY_IS_EQUAL(lt(i,j,k), mat1(i,j,k) < mat2(i,j,k)); in test_orderings()
38 VERIFY_IS_EQUAL(le(i,j,k), mat1(i,j,k) <= mat2(i,j,k)); in test_orderings()
39 VERIFY_IS_EQUAL(gt(i,j,k), mat1(i,j,k) > mat2(i,j,k)); in test_orderings()
40 VERIFY_IS_EQUAL(ge(i,j,k), mat1(i,j,k) >= mat2(i,j,k)); in test_orderings()
[all …]
/aosp_15_r20/external/pytorch/torch/_inductor/kernel/
H A Dmm.py144 def bias_addmm(inp, mat1, mat2, *, out=None, alpha=1, beta=1): argument
151 return torch.addmm(inp[0], mat1, mat2, out=out, alpha=alpha, beta=beta)
152 return torch.addmm(inp, mat1, mat2, out=out, alpha=alpha, beta=beta)
159 def tuned_mm(mat1, mat2, *, layout=None): argument
160 m, n, k, layout, mat1, mat2 = mm_args(mat1, mat2, layout=layout)
171 [aten_mm.bind((mat1, mat2), aten_layout)] if use_aten_gemm_kernels() else []
173 static_shape, is_nonzero = _is_static_problem([mat1, mat2], layout)
178 input_nodes=(mat1, mat2),
183 CUTLASS3xGemmTemplate.add_cutlass_gemm_choices(choices, layout, [mat1, mat2])
186 CKGemmTemplate.add_ck_gemm_choices(choices, layout, [mat1, mat2])
[all …]
H A Dbmm.py107 def tuned_bmm(mat1, mat2, *, layout=None): argument
108 if all(x.get_device().type == "cpu" for x in [mat1, mat2]):
110 if mat1.get_size()[1] == 1 or mat2.get_size()[2] == 1:
111 mat1 = L.unsqueeze(mat1, -1)
113 return L.sum_(L.mul(mat1, mat2), axis=2)
138 if is_valid_to_require_contiguous(mat1):
140 mat1 = may_require_contiguous(mat1, meta_mat1)
145 m, n, k, layout, mat1, mat2 = mm_args(mat1, mat2, layout=layout)
148 choices = [aten_bmm.bind((mat1, mat2), layout)] if use_aten_gemm_kernels() else []
153 input_nodes=(mat1, mat2),
[all …]
/aosp_15_r20/external/pytorch/torch/_inductor/fx_passes/
H A Dpad_mm.py85 mat1: Tensor, mat2: Tensor, input: Optional[Tensor] = None
113 and check_device(mat1, mat2)
114 and check_dtype(mat1, mat2)
115 and all(valid_shape_and_stride(t) for t in (mat1, mat2, input))
139 input: Tensor, mat1: Tensor, mat2: Tensor, beta: float, alpha: float
141 return aten.addmm(input, mat1, mat2, beta=beta, alpha=alpha)
145 mat1, mat2, input = fetch_fake_tensors(match, ("mat1", "mat2", "input"))
146 return should_pad_common(mat1, mat2, input) and should_pad_bench(
147 match, mat1, mat2, torch.ops.aten.addmm, input=input
153 mat1: Tensor,
[all …]
H A Ddecompose_mem_bound_mm.py42 def should_decompose_bmm(mat1, mat2) -> bool: argument
43 if is_node_meta_valid(mat1) and is_node_meta_valid(mat2):
44 mat1 = mat1.meta["val"]
48 if not check_device(mat1, mat2):
51 if len(mat1.shape) != 3 or len(mat2.shape) != 3:
53 if mat1.shape[0] < min_first_dimension_decomposition:
56 if (mat1.shape[1] < max_other_dimention_decomposition) + (
57 mat1.shape[2] < max_other_dimention_decomposition
63 def should_decompose_mm(mat1, mat2) -> bool: argument
64 if is_node_meta_valid(mat1) and is_node_meta_valid(mat2):
[all …]
/aosp_15_r20/external/eigen/doc/
H A DQuickReference.dox330 mat3 = mat1 + mat2; mat3 += mat1;
331 mat3 = mat1 - mat2; mat3 -= mat1;\endcode
335 mat3 = mat1 * s1; mat3 *= s1; mat3 = s1 * mat1;
336 mat3 = mat1 / s1; mat3 /= s1;\endcode
340 col2 = mat1 * col1;
341 row2 = row1 * mat1; row1 *= mat1;
342 mat3 = mat1 * mat2; mat3 *= mat1; \endcode
346 mat1 = mat2.transpose(); mat1.transposeInPlace();
347 mat1 = mat2.adjoint(); mat1.adjointInPlace();
433 mat1.real()
[all …]
/aosp_15_r20/external/pytorch/aten/src/ATen/native/mkldnn/
H A DMatmul.cpp13 const Tensor &mat1, in mkldnn_matmul() argument
22 const Tensor& mat1, in use_mkldnn_bf16_matmul() argument
29 const Tensor& mat1, in use_mkldnn_fp16_matmul() argument
68 const Tensor& mat1, in use_mkldnn_bf32_matmul() argument
75 const Tensor& mat1, in use_mkldnn_matmul() argument
82 const Tensor &mat1, in mkldnn_matmul_i8i8i32() argument
228 const Tensor &mat1, in mkldnn_matmul() argument
233 TORCH_CHECK((mat1.dim() == 2 && mat2.dim() == 2) || // aten::addmm in mkldnn_matmul()
234 (mat1.dim() == 3 && mat2.dim() == 3) || // aten::bmm, aten::baddbmm in mkldnn_matmul()
235 (mat1.dim() == 2 && mat2.dim() == 1) || // aten::mv in mkldnn_matmul()
[all …]
/aosp_15_r20/external/pytorch/aten/src/ATen/native/sparse/
H A DSparseBlas.cpp112 * `mat1` - [in] dense Tensor A of size m × k.
119 const Tensor& mat1, in sparse_sampled_addmm_out_sparse_csr_cpu() argument
124 at::native::sparse::sparse_sampled_addmm_check_inputs(self, mat1, mat2, beta, alpha, result); in sparse_sampled_addmm_out_sparse_csr_cpu()
131 // We allow self to be a single matrix when mat1 and mat2 are batched in sparse_sampled_addmm_out_sparse_csr_cpu()
132 auto result_sizes = DimVector(mat1.sizes().slice(0, mat1.dim() - 2)); in sparse_sampled_addmm_out_sparse_csr_cpu()
139 if (mat1.numel() == 0 || mat2.numel() == 0 || result._nnz() == 0) { in sparse_sampled_addmm_out_sparse_csr_cpu()
148 sampled_addmm_sparse_csr_stub(kCPU, mat1.contiguous(), mat2_t, beta, alpha, result); in sparse_sampled_addmm_out_sparse_csr_cpu()
155 const Tensor& mat1, in sparse_sampled_addmm_sparse_csr_cpu() argument
160 at::native::sparse_sampled_addmm_out_sparse_csr_cpu(self, mat1, mat2, beta, alpha, result); in sparse_sampled_addmm_sparse_csr_cpu()
168 const Tensor& mat1, in sparse_sampled_addmm_check_inputs() argument
[all …]
/aosp_15_r20/external/deqp-deps/glslang/Test/baseResults/
Dhlsl.matpack-pragma.frag.out17 0:31 mat1: direct index for structure (layout( row_major) temp 4X4 matrix of flo…
18 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
19 …mat1, layout( column_major) temp 4X4 matrix of float mat2, layout( column_major) temp 4X4 matrix o…
28 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
29 …mat1, layout( column_major) temp 4X4 matrix of float mat2, layout( column_major) temp 4X4 matrix o…
38 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
39 …mat1, layout( column_major) temp 4X4 matrix of float mat2, layout( column_major) temp 4X4 matrix o…
47 0:32 mat1: direct index for structure (layout( row_major) temp 4X4 matrix of float)
48 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
49 …mat1, layout( column_major) temp 4X4 matrix of float mat2, layout( column_major) temp 4X4 matrix o…
[all …]
Dhlsl.matpack-1.frag.out12 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
13 …mat1, layout( column_major) temp 4X4 matrix of float mat2, temp 4-component vector of float vec1,…
18 0:24 mat1: direct index for structure (layout( row_major) temp 4X4 matrix of float)
19 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
20 …mat1, layout( column_major) temp 4X4 matrix of float mat2, temp 4-component vector of float vec1,…
27 … std140) uniform structure{layout( column_major) temp 4X4 matrix of float mat1, temp 4-component …
28 …mat1, layout( column_major) temp 4X4 matrix of float mat2, temp 4-component vector of float vec1,…
33 0:25 mat1: direct index for structure (layout( column_major) temp 4X4 matrix of float)
34 … std140) uniform structure{layout( column_major) temp 4X4 matrix of float mat1, temp 4-component …
35 …mat1, layout( column_major) temp 4X4 matrix of float mat2, temp 4-component vector of float vec1,…
[all …]
/aosp_15_r20/external/angle/third_party/glslang/src/Test/baseResults/
H A Dhlsl.matpack-pragma.frag.out17 0:31 mat1: direct index for structure (layout( row_major) temp 4X4 matrix of flo…
18 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
19 …mat1, layout( column_major) temp 4X4 matrix of float mat2, layout( column_major) temp 4X4 matrix o…
28 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
29 …mat1, layout( column_major) temp 4X4 matrix of float mat2, layout( column_major) temp 4X4 matrix o…
38 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
39 …mat1, layout( column_major) temp 4X4 matrix of float mat2, layout( column_major) temp 4X4 matrix o…
47 0:32 mat1: direct index for structure (layout( row_major) temp 4X4 matrix of float)
48 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
49 …mat1, layout( column_major) temp 4X4 matrix of float mat2, layout( column_major) temp 4X4 matrix o…
[all …]
H A Dhlsl.matpack-1.frag.out12 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
13 …mat1, layout( column_major) temp 4X4 matrix of float mat2, temp 4-component vector of float vec1,…
18 0:24 mat1: direct index for structure (layout( row_major) temp 4X4 matrix of float)
19 …jor std140) uniform structure{layout( row_major) temp 4X4 matrix of float mat1, layout( column_maj…
20 …mat1, layout( column_major) temp 4X4 matrix of float mat2, temp 4-component vector of float vec1,…
27 … std140) uniform structure{layout( column_major) temp 4X4 matrix of float mat1, temp 4-component …
28 …mat1, layout( column_major) temp 4X4 matrix of float mat2, temp 4-component vector of float vec1,…
33 0:25 mat1: direct index for structure (layout( column_major) temp 4X4 matrix of float)
34 … std140) uniform structure{layout( column_major) temp 4X4 matrix of float mat1, temp 4-component …
35 …mat1, layout( column_major) temp 4X4 matrix of float mat2, temp 4-component vector of float vec1,…
[all …]
/aosp_15_r20/external/executorch/backends/vulkan/runtime/graph/ops/impl/
H A DQuantizedLinear.cpp20 const ValueRef mat1, in check_q_8w_linear_args() argument
24 std::vector<int64_t> mat1_sizes = graph.sizes_of(mat1); in check_q_8w_linear_args()
32 VK_CHECK_COND(graph.packed_dim_of(mat1) == graph.packed_dim_of(out)); in check_q_8w_linear_args()
47 vTensorPtr mat1 = graph->get_tensor(args[1].refs[0]); in resize_q_8w_linear_node() local
50 const int out_cols = utils::val_at(-2, mat1->sizes()); in resize_q_8w_linear_node()
54 if (mat1->sizes().size() == 2) { in resize_q_8w_linear_node()
59 new_out_sizes.at(0) = mat1->sizes().at(0); in resize_q_8w_linear_node()
69 const ValueRef mat1, in add_q_8w_linear_node() argument
74 ValueRef mat1_W_packed = mat1; in add_q_8w_linear_node()
77 graph.packed_dim_of(mat1) != WHCN::kWidthDim) { in add_q_8w_linear_node()
[all …]
H A DMatMul.cpp23 const ValueRef mat1, in check_matmul_args() argument
26 std::vector<int64_t> mat1_sizes = graph.sizes_of(mat1); in check_matmul_args()
32 VK_CHECK_COND(graph.packed_dim_of(mat1) == graph.packed_dim_of(out)); in check_matmul_args()
42 vTensorPtr mat1 = graph->get_tensor(args[1].refs[0]); in resize_matmul_node() local
47 const int out_cols = utils::val_at(-2, mat1->sizes()); in resize_matmul_node()
52 std::vector<int64_t> new_out_sizes(mat1->sizes()); in resize_matmul_node()
61 const ValueRef mat1, in add_matmul_naive_buffer_node() argument
92 {{mat1, mat2}, vkapi::MemoryAccessType::READ}}, in add_matmul_naive_buffer_node()
97 graph.sizes_ubo(mat1), in add_matmul_naive_buffer_node()
98 graph.strides_ubo(mat1), in add_matmul_naive_buffer_node()
[all …]
H A DLinear.cpp24 const ValueRef mat1, in check_addmm_args() argument
33 std::vector<int64_t> mat1_sizes = graph.sizes_of(mat1); in check_addmm_args()
39 VK_CHECK_COND(graph.packed_dim_of(mat1) == graph.packed_dim_of(out)); in check_addmm_args()
58 vTensorPtr mat1 = graph->get_tensor(args[1].refs[0]); in resize_addmm_node() local
64 const int out_cols = utils::val_at(-2, mat1->sizes()); in resize_addmm_node()
69 if (mat1->sizes().size() == 2) { in resize_addmm_node()
74 new_out_sizes.at(0) = mat1->sizes().at(0); in resize_addmm_node()
90 const ValueRef mat1, in add_addmm_naive_node() argument
117 {{mat1, mat2, self}, vkapi::MemoryAccessType::READ}}, in add_addmm_naive_node()
122 graph.sizes_ubo(mat1), in add_addmm_naive_node()
[all …]
/aosp_15_r20/external/pytorch/aten/src/ATen/native/sparse/cuda/
H A DSparseBlas.cpp33 * `mat1` - [in] dense Tensor A of size m × k.
40 const Tensor& mat1, in sparse_sampled_addmm_out_sparse_csr_cuda() argument
46 self, mat1, mat2, beta, alpha, result); in sparse_sampled_addmm_out_sparse_csr_cuda()
49 // We allow self to be a single matrix when mat1 and mat2 are batched in sparse_sampled_addmm_out_sparse_csr_cuda()
50 auto result_sizes = DimVector(mat1.sizes().slice(0, mat1.dim() - 2)); in sparse_sampled_addmm_out_sparse_csr_cuda()
58 if (mat1.numel() == 0 || mat2.numel() == 0) { in sparse_sampled_addmm_out_sparse_csr_cuda()
63 sparse::impl::cuda::sampled_addmm_out_sparse_csr(mat1, mat2, beta, alpha, result); in sparse_sampled_addmm_out_sparse_csr_cuda()
69 const Tensor& mat1, in sparse_sampled_addmm_sparse_csr_cuda() argument
74 at::native::sparse_sampled_addmm_out_sparse_csr_cuda(self, mat1, mat2, beta, alpha, result); in sparse_sampled_addmm_sparse_csr_cuda()
78 // result = beta * self + alpha * (mat1 @ mat2)
[all …]
H A DSparseBlasImpl.cpp73 const at::sparse_csr::SparseCsrTensor& mat1, in addmm_out_legacy() argument
78 TORCH_INTERNAL_ASSERT_DEBUG_ONLY(mat1.is_sparse_csr()); in addmm_out_legacy()
79 auto nnz = mat1._nnz(); in addmm_out_legacy()
80 auto m = mat1.size(0); in addmm_out_legacy()
81 auto k = mat1.size(1); in addmm_out_legacy()
83 auto crow_indices = mat1.crow_indices().to(kInt); in addmm_out_legacy()
84 auto col_indices = mat1.col_indices().to(kInt); in addmm_out_legacy()
85 auto values = mat1.values(); in addmm_out_legacy()
464 const at::sparse_csr::SparseCsrTensor& mat1, in block_sparse_mm() argument
469 TORCH_INTERNAL_ASSERT_DEBUG_ONLY(mat1.layout() == kSparseBsr); in block_sparse_mm()
[all …]
/aosp_15_r20/external/vixl/examples/aarch64/
H A Dneon-matrix-multiply.cc63 // mat1 -> x1 in GenerateNEONMatrixMultiply()
117 float mat1[kLength], mat2[kLength], output[kLength]; in main() local
124 // float mat1[kLength] = { 1.0f, 52.03f, 4.43f, ... }; in main()
127 mat1[0] = 1.0f; in main()
128 mat1[4] = 2.0f; in main()
129 mat1[8] = 3.0f; in main()
130 mat1[12] = 4.0f; in main()
131 mat1[1] = 52.03f; in main()
132 mat1[5] = 12.24f; in main()
133 mat1[9] = 53.56f; in main()
[all …]
/aosp_15_r20/external/pytorch/aten/src/ATen/native/mkl/
H A DSparseBlasImpl.cpp288 * `mat1` - Sparse CSR Tensor storing m x k matrix A.
294 const Tensor& mat1, in addmm_sparse_result() argument
314 if (mat1._nnz() == 0 || mat2._nnz() == 0) { in addmm_sparse_result()
325 at::mkl::sparse::MklSparseCsrDescriptor<scalar_t>(mat1); in addmm_sparse_result()
353 * `mat1` - Tensor storing m x k matrix A.
359 const Tensor& mat1, in addmm_out_sparse_csr() argument
365 mat1.dim() == 2 && mat2.dim() == 2 && result.dim() == 2); in addmm_out_sparse_csr()
367 !((mat1.layout() == kStrided) && (mat2.layout() == kStrided) && in addmm_out_sparse_csr()
371 // Layout checks are nested mat1, mat2, result in addmm_out_sparse_csr()
376 if (mat1.layout() == kStrided) { in addmm_out_sparse_csr()
[all …]
/aosp_15_r20/external/pytorch/test/distributed/_tensor/
H A Dtest_common_rules.py41 mat1, mat2 = [-1, -1], [-1, 0]
46 mesh, mat1, [], tensor_meta=mat1_tensor_meta
59 mat1, mat2 = [0, -1], [-1, -1]
61 mesh, mat1, [], tensor_meta=mat1_tensor_meta
74 mat1, mat2 = [-1, 0], [0, -1]
76 mesh, mat1, [], tensor_meta=mat1_tensor_meta
95 mat1 = [0, -1]
97 mesh, mat1, [], tensor_meta=mat1_tensor_meta
108 mat1 = [-1, 0, -1]
110 mesh, mat1, [], tensor_meta=mat1_tensor_meta
[all …]
/aosp_15_r20/external/eigen/test/
H A Devaluators.cpp250 MatrixXd mat1(6,6), mat2(6,6); in EIGEN_DECLARE_TEST() local
251 VERIFY_IS_APPROX_EVALUATOR(mat1, MatrixXd::Identity(6,6)); in EIGEN_DECLARE_TEST()
252 VERIFY_IS_APPROX_EVALUATOR(mat2, mat1); in EIGEN_DECLARE_TEST()
253 copy_using_evaluator(mat2.transpose(), mat1); in EIGEN_DECLARE_TEST()
254 VERIFY_IS_APPROX(mat2.transpose(), mat1); in EIGEN_DECLARE_TEST()
262 VERIFY_IS_APPROX_EVALUATOR(mat2, mat1); in EIGEN_DECLARE_TEST()
332 mat1.setRandom(); in EIGEN_DECLARE_TEST()
335 copy_using_evaluator(matXcd.real(), mat1); in EIGEN_DECLARE_TEST()
337 matXcd_ref.real() = mat1; in EIGEN_DECLARE_TEST()
355 VERIFY_IS_APPROX_EVALUATOR(vec1, mat1.rowwise().sum()); in EIGEN_DECLARE_TEST()
[all …]
/aosp_15_r20/external/pytorch/aten/src/ATen/native/cuda/
H A DBlas.cpp97 cublasCommonArgs(const Tensor& mat1, const Tensor& mat2, Tensor& c) { in cublasCommonArgs()
100 …mata = prepare_matrix_for_cublas(transpose_result ? mat2 : mat1, transpose_mat1, transpose_result); in cublasCommonArgs()
101 …matb = prepare_matrix_for_cublas(transpose_result ? mat1 : mat2, transpose_mat2, transpose_result); in cublasCommonArgs()
102 auto mat1_sizes = mat1.sizes(); in cublasCommonArgs()
256 Tensor& addmm_out_cuda_impl(Tensor& result, const Tensor& self, const Tensor& mat1, const Tensor& m… in addmm_out_cuda_impl() argument
260 TORCH_CHECK(mat1.dim() == 2 && mat2.dim() == 2, "tensors must be 2-D"); in addmm_out_cuda_impl()
262 mat1.dtype() == mat2.dtype(), in addmm_out_cuda_impl()
263 "expected mat1 and mat2 to have the same dtype, but got: ", mat1.dtype(), " != ", mat2.dtype() in addmm_out_cuda_impl()
266 TensorArg targs[]{{result, "out", 0}, {self, "self", 1}, {mat1, "mat1", 2}, {mat2, "mat2", 3}}; in addmm_out_cuda_impl()
269 IntArrayRef mat1_sizes = mat1.sizes(); in addmm_out_cuda_impl()
[all …]

12345678