Home
last modified time | relevance | path

Searched refs:input_shape_ (Results 1 – 25 of 123) sorted by relevance

12345

/third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/gpu/arrays/
Dslice_gpu_kernel.h47 size_t input_rank = input_shape_.size(); in Launch()
50 …Slice1DKernel(begin_[0], size_[0], input_shape_[0], input, output, reinterpret_cast<cudaStream_t>(… in Launch()
53 …Slice2DKernel(begin_[0], begin_[1], size_[0], size_[1], input_shape_[0], input_shape_[1], input, o… in Launch()
57 …el(begin_[0], begin_[1], begin_[2], size_[0], size_[1], size_[2], input_shape_[0], input_shape_[1], in Launch()
58 input_shape_[2], input, output, reinterpret_cast<cudaStream_t>(stream_ptr)); in Launch()
62 … input_shape_[0], input_shape_[1], input_shape_[2], input_shape_[3], input, output, in Launch()
67 … size_[4], input_shape_[0], input_shape_[1], input_shape_[2], input_shape_[3], input_shape_[4], in Launch()
72 … size_[3], size_[4], size_[5], input_shape_[0], input_shape_[1], input_shape_[2], input_shape_[3], in Launch()
73 … input_shape_[4], input_shape_[5], input, output, reinterpret_cast<cudaStream_t>(stream_ptr)); in Launch()
77 … size_[2], size_[3], size_[4], size_[5], size_[6], input_shape_[0], input_shape_[1], in Launch()
[all …]
Din_top_k_gpu_kernel.h83 …(casted_float32_input, targets_device, output_device, top_k_output_device_float32, input_shape_[0], in Launch()
84 input_shape_[1], k_, reinterpret_cast<cudaStream_t>(stream_ptr)); in Launch()
90 … CalInTopK(predictions_device, targets_device, output_device, top_k_output_device, input_shape_[0], in Launch()
91 input_shape_[1], k_, reinterpret_cast<cudaStream_t>(stream_ptr)); in Launch()
111 input_shape_ = AnfAlgo::GetInputDeviceShape(kernel_node, 0); in Init()
112 if (input_shape_.size() < 2) { in Init()
114 << input_shape_.size(); in Init()
116 is_null_input_ = CHECK_NULL_INPUT(input_shape_); in Init()
122 input_rank_ = input_shape_.size(); in Init()
125 input_size_ *= input_shape_[i]; in Init()
[all …]
Dslice_grad_gpu_kernel.h48 …egin_[0], begin_[1], begin_[2], begin_[3], size_[0], size_[1], size_[2], size_[3], input_shape_[0], in Launch()
49 input_shape_[1], input_shape_[2], input_shape_[3], dy, dx, in Launch()
64 input_shape_.push_back(static_cast<size_t>(x)); in Init()
66 for (auto i = input_shape_.size(); i < 4; i++) { in Init()
67 (void)input_shape_.insert(input_shape_.begin(), 1); in Init()
82 ShapeNdTo4d(input_shape, &input_shape_); in Init()
95 input_size_ = input_shape_[0] * input_shape_[1] * input_shape_[2] * input_shape_[3] * sizeof(T); in Init()
124 if (begin_[i] < 0 && i < input_shape_.size()) { in CalcBeginAndSize()
125 begin_[i] = begin_[i] + input_shape_[i]; in CalcBeginAndSize()
129 if (size_[i] < 0 && i < input_shape_.size()) { in CalcBeginAndSize()
[all …]
Dtensor_copy_slices_gpu_kernel.h74 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(kernel_node, 0); in Init()
76 is_null_input_ = CHECK_NULL_INPUT(input_shape_) || CHECK_NULL_INPUT(update_shape); in Init()
82 if (input_shape_.size() > kMaxDims) { in Init()
84 << input_shape_.size(); in Init()
92 if (begin_.size() > input_shape_.size()) { in Init()
95 << ", the rank of input: " << input_shape_.size(); in Init()
100 output_shape_ = input_shape_; in Init()
129 for (size_t i = 0; i < input_shape_.size(); i++) { in GetSize()
130 input_size_ *= input_shape_[i]; in GetSize()
153 int64_t dim = input_shape_[i]; in FillEmptyDims()
[all …]
Dbroadcast_to_gpu_kernel.h46 …BroadcastTo(input_shape_[0], input_shape_[1], input_shape_[2], input_shape_[3], output_shape_[0], … in Launch()
70 input_shape_[i + offset] = input_shapes[i]; in Init()
83 …input_size_list_.push_back(input_shape_[0] * input_shape_[1] * input_shape_[2] * input_shape_[3] *… in InitSizeLists()
88 size_t input_shape_[4] = {1, 1, 1, 1};
Dtile_gpu_kernel.h48 … cudaMemcpyAsync(input_shape_ptr, &input_shape_[0], input_shape_.size() * sizeof(size_t), in Launch()
71 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(kernel_node, 0); in Init()
73 is_null_input_ = CHECK_NULL_INPUT(input_shape_) || CHECK_NULL_INPUT(output_shape_); in Init()
84 for (size_t i = 0; i < input_shape_.size(); i++) { in Init()
85 input_size_ *= input_shape_[i]; in Init()
98 …filling_value = static_cast<int64_t>(multiples.size()) - static_cast<int64_t>(input_shape_.size()); in Init()
100 (void)input_shape_.insert(input_shape_.begin(), LongToSize(filling_value), 1); in Init()
110 input_shape_.clear(); in ResetResource()
120 workspace_size_list_.push_back(input_shape_.size() * sizeof(size_t)); in InitSizeLists()
130 std::vector<size_t> input_shape_; variable
Dstrided_slice_gpu_kernel.h47 StridedSlice(input_shape_, begin_, strides_, output_shape_, input, output, in Launch()
53 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(kernel_node, 0); in Init()
54 null_output_ = CHECK_NULL_INPUT(input_shape_); in Init()
60 if (input_shape_.size() > MAX_DIMS) { in Init()
62 << input_shape_.size(); in Init()
74 for (size_t i = 0; i < input_shape_.size(); i++) { in InitSizeLists()
75 size *= input_shape_[i]; in InitSizeLists()
Dreverse_sequence_gpu_kernel.h60 … cudaMemcpyAsync(input_shape_ptr, &input_shape_[0], input_shape_.size() * sizeof(size_t), in Launch()
81 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(kernel_node, 0); in Init()
83 is_null_input_ = CHECK_NULL_INPUT(input_shape_) || CHECK_NULL_INPUT(seq_len_shape); in Init()
89 if (input_shape_.size() < 1) { in Init()
91 << input_shape_.size(); in Init()
94 shape_size_ = input_shape_.size(); // required for calls in Init()
96 input_size_ *= input_shape_[i]; in Init()
128 std::vector<size_t> input_shape_; variable
Dstrided_slice_gpu_common.h47 if (i >= input_shape_.size()) { in FillEmptyDims()
48 input_shape_.push_back(1); in FillEmptyDims()
52 int64_t dim = input_shape_[i]; in FillEmptyDims()
59 int64_t dim = input_shape_[i]; in FillEmptyDims()
62 end_.push_back(i < input_shape_.size() ? input_shape_[i] : 1); in FillEmptyDims()
84 end_[j] = input_shape_[j]; in ParseMasks()
93 end_[k] = input_shape_[k]; in ParseMasks()
103 end_[l] = input_shape_[l]; in ParseMasks()
158 std::vector<size_t> input_shape_; variable
Dresize_nearest_neighbor_grad_gpu_kernel.h51 float h_scale = Scaling(output_shape_[2], input_shape_[2], align_corners_); in Launch()
52 float w_scale = Scaling(output_shape_[3], input_shape_[3], align_corners_); in Launch()
53 …ResizeNearestNeighborGrad(input_size, input, input_shape_[0], input_shape_[1], input_shape_[2], in… in Launch()
95 input_shape_.push_back(input_shape[i]); in Init()
128 std::vector<int> input_shape_; variable
Dresize_nearest_neighbor_gpu_kernel.h51 float h_scale = Scaling(input_shape_[2], output_shape_[2], align_corners_); in Launch()
52 float w_scale = Scaling(input_shape_[3], output_shape_[3], align_corners_); in Launch()
53 …CalResizeNearestNeighbor(size, input, input_shape_[0], input_shape_[1], input_shape_[2], input_sha… in Launch()
91 input_shape_.push_back(input_shape[i]); in Init()
120 std::vector<int> input_shape_; variable
/third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/cpu/
Dbias_add_cpu_kernel.cc33 input_shape_ = AnfAlgo::GetInputDeviceShape(kernel_node, 0); in InitKernel()
35 data_shape_ = input_shape_.size(); in InitKernel()
36 if (input_shape_.size() < kBiasAddMinDim || input_shape_.size() > kBiasAddMaxDim) { in InitKernel()
39 << input_shape_.size(); in InitKernel()
44 if (input_shape_[1] != bias_shape_[0]) { in InitKernel()
46 << input_shape_[1] << "]"; in InitKernel()
58 if (input_shape_.size() > 2) { in Launch()
60 for (size_t i = 2; i < input_shape_.size(); ++i) { in Launch()
61 hw_size *= input_shape_[i]; in Launch()
64 size_t c_size = input_shape_[1]; in Launch()
[all …]
Dbias_add_grad_cpu_kernel.cc31 input_shape_ = AnfAlgo::GetInputDeviceShape(kernel_node, 0); in InitKernel()
32 if (input_shape_.size() < 2) { in InitKernel()
34 << input_shape_.size(); in InitKernel()
45 if (input_shape_.size() > 2) { in Launch()
47 for (size_t i = 2; i < input_shape_.size(); ++i) { in Launch()
48 hw_size *= input_shape_[i]; in Launch()
51 size_t c_size = input_shape_[1]; in Launch()
54 for (size_t n = 0; n < input_shape_[0]; ++n) { in Launch()
61 } else if (input_shape_.size() == 2) { in Launch()
64 …ReduceSumDim2Axis0(end - start, input_shape_[1], input_shape_[0], input_addr + start, output_addr … in Launch()
[all …]
Dsplit_cpu_kernel.cc38 (void)std::transform(input_shape.begin(), input_shape.end(), std::back_inserter(input_shape_), in InitKernel()
40 if (input_shape_.size() < 1 || input_shape_.size() > SPLIT_STRIDES_SIZE) { in InitKernel()
42 << input_shape_.size(); in InitKernel()
68 param.strides_[input_shape_.size() - 1] = 1; in LaunchSplit()
69 for (int i = SizeToInt(input_shape_.size()) - 2; i >= 0; i--) { // from -2 to 0 dim in LaunchSplit()
70 param.strides_[i] = param.strides_[i + 1] * input_shape_[i + 1]; in LaunchSplit()
74 int split_size = input_shape_[param.split_dim_] / SizeToInt(output_num_); in LaunchSplit()
80 param.split_count_ *= input_shape_[i]; in LaunchSplit()
83 …(void)DoSplit(input, reinterpret_cast<void **>(output), &input_shape_[0], SizeToInt(start), SizeTo… in LaunchSplit()
106 int64_t dims = SizeToLong(input_shape_.size()); in CheckParam()
[all …]
Dbroadcast_to_cpu_kernel.cc31 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(kernel_node, 0); in InitKernel()
33 size_t input_shape_size = input_shape_.size(); in InitKernel()
36 MS_LOG(EXCEPTION) << "Cannot broadcast input tensor with shape " << input_shape_ in InitKernel()
40 …MS_LOG(EXCEPTION) << "Cannot broadcast input tensor with shape " << input_shape_ << " to a shape "… in InitKernel()
45 if (input_shape_[i] != output_shape_[i + offset] && input_shape_[i] != 1) { in InitKernel()
46 … MS_LOG(EXCEPTION) << "Cannot broadcast input tensor with shape " << input_shape_ << " to a shape " in InitKernel()
52 shape_info_.input_shape_[i] = SizeToInt(input_shape_[i]); in InitKernel()
80 …MS_LOG(EXCEPTION) << "Broadcast tensor with shape " << input_shape_ << " to shape " << output_shap… in Launch()
Dstridedslice_cpu_kernel.cc46 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(kernel_node, 0); in InitKernel()
48 if (input_shape_.size() > DIMENSION_8D || input_shape_.empty()) { in InitKernel()
49 …ION) << "StridedSlice only support 1D to 8D input tensor, but got " << input_shape_.size() << "D."; in InitKernel()
55 …if (begin.size() != end.size() || begin.size() != stride.size() || begin.size() > input_shape_.siz… in InitKernel()
77 if (input_shape_.size() == output_shape_.size()) { in MatchParallelPattern()
79 for (size_t i = 0; i < input_shape_.size(); ++i) { in MatchParallelPattern()
80 if (input_shape_[i] != output_shape_[i]) { in MatchParallelPattern()
94 …std::accumulate(input_shape_.begin(), input_shape_.begin() + split_axis_, size_t(1), std::multipli… in InitParallelParam()
96 …std::accumulate(input_shape_.begin() + split_axis_ + 1, input_shape_.end(), size_t(1), std::multip… in InitParallelParam()
131 dim_len = SizeToInt(input_shape_[i]); in InitSliceParam()
[all …]
Dmultinomial_cpu_kernel.cc25 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(kernel_node, 0); in InitKernel()
27 if (input_shape_.size() == 1) { in InitKernel()
28 workspace_size_list_.push_back(input_shape_[0] * sizeof(float)); in InitKernel()
29 } else if (input_shape_.size() == 2) { in InitKernel()
30 workspace_size_list_.push_back(input_shape_[1] * sizeof(float)); in InitKernel()
65 if (input_shape_.size() == 2) { in Launch()
66 num_row = input_shape_[0]; in Launch()
68 int num_col = input_shape_[input_shape_.size() - 1]; in Launch()
Dl2_normalize_cpu_kernel.cc35 input_shape_ = AnfAlgo::GetInputDeviceShape(kernel_node, 0); in InitKernel()
38 int dims = SizeToInt(input_shape_.size()); in InitKernel()
46 axis_ += SizeToInt(input_shape_.size()); in InitKernel()
55 std::vector<size_t> axes(input_shape_.size()); in CalcDenominator()
63 stride *= input_shape_[i]; in CalcDenominator()
68 std::vector<size_t> transpose_shape(input_shape_.size()); in CalcDenominator()
70 transpose_shape[i] = input_shape_[axes[i]]; in CalcDenominator()
73 TransposeIterator tran_base_iter(std::move(transpose_shape), std::move(axes), input_shape_); in CalcDenominator()
100 BroadcastIterator broad_base_iter(input_shape_, reduce_shape, output_shape_); in CalcOutput()
135 int dims = SizeToInt(input_shape_.size()); in Launch()
[all …]
Dgather_cpu_kernel.cc35 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(kernel_node, 0); in InitKernel()
38 if (input_shape_.size() > kGatherInputParamsMaxDim) { in InitKernel()
39 …MS_LOG(EXCEPTION) << "Input dims is " << input_shape_.size() << ", but GatherV2CPUKernel olny supp… in InitKernel()
60 int dims = SizeToInt(input_shape_.size()); in Launch()
79 outer_size *= input_shape_.at(i); in ParallelRun()
81 for (size_t i = axis + 1; i < input_shape_.size(); ++i) { in ParallelRun()
82 inner_size *= input_shape_.at(i); in ParallelRun()
88 auto limit = input_shape_.at(axis); in ParallelRun()
/third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/gpu/nn/
Dbce_with_logits_loss_kernel.h52 … cudaMemcpyAsync(input_shape, &input_shape_[0], input_shape_.size() * sizeof(size_t), in Launch()
64 …CalBCEWithLogitsLoss(input_size_, predict, target, input_shape, input_shape_.size(), weight, weigh… in Launch()
80 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(kernel_node, 0); in Init()
84 …CHECK_NULL_INPUT(input_shape_) || CHECK_NULL_INPUT(weight_shape_) || CHECK_NULL_INPUT(pos_weight_s… in Init()
90 if (input_shape_.size() < 1) { in Init()
92 << input_shape_.size(); in Init()
103 if (input_shape_.size() > MAX_LOGITS_DIMENSION) { in Init()
104 MS_LOG(EXCEPTION) << "Input dimension is " << input_shape_.size() in Init()
107 for (size_t i = 0; i < input_shape_.size(); i++) { in Init()
108 input_size_ *= input_shape_[i]; in Init()
[all …]
Dinstance_norm_gpu_kernel.h80 size_t N = input_shape_[0]; in Launch()
81 size_t C = input_shape_[1]; in Launch()
114 input_shape_ = AnfAlgo::GetInputDeviceShape(kernel_node, 0); in Init()
115 if (input_shape_.size() != 4) { in Init()
116 …MS_LOG(EXCEPTION) << "tensor shape is " << input_shape_.size() << ", InstanceNormGpuKernel should … in Init()
118 is_null_input_ = CHECK_NULL_INPUT(input_shape_); in Init()
124 CheckTensorSize({input_shape_}); in Init()
163 input_size_list_.push_back(input_shape_[1]); // gamma in InitSizeLists()
164 input_size_list_.push_back(input_shape_[1]); // beta in InitSizeLists()
165 input_size_list_.push_back(input_shape_[1]); // mean in InitSizeLists()
[all …]
Dinstance_norm_grad_gpu_kernel.h81 size_t N = input_shape_[0]; in Launch()
82 size_t C = input_shape_[1]; in Launch()
119 input_shape_ = AnfAlgo::GetInputDeviceShape(kernel_node, 0); in Init()
120 if (input_shape_.size() != 4) { in Init()
121 …MS_LOG(EXCEPTION) << "tensor shape is " << input_shape_.size() << ", InstanceNormGradGpuKernel sho… in Init()
123 is_null_input_ = CHECK_NULL_INPUT(input_shape_); in Init()
129 CheckTensorSize({input_shape_}); in Init()
160 input_size_list_.push_back(input_shape_[1]); in InitSizeLists()
184 int channel = SizeToInt(input_shape_[0]) * SizeToInt(input_shape_[1]); in SetTensorDescriptor()
185 int height = SizeToInt(input_shape_[2]); in SetTensorDescriptor()
[all …]
/third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/cpu/ps/
Dembedding_look_up_ps_kernel.cc41 input_shape_ = *(shape_vec[0]); in InitKernel()
42 if (input_shape_.empty()) { in InitKernel()
45 first_dim_size_ = input_shape_[0]; in InitKernel()
46 for (size_t i = 1; i < input_shape_.size(); ++i) { in InitKernel()
47 outer_dim_size_ *= input_shape_[i]; in InitKernel()
58 …offset += Util::LocalShard(SizeToLong(input_shape_[kAxis]), SizeToLong(i), SizeToLong(pserver_num_… in InitKernel()
63 Shard(&input_shape_, kAxis); in InitKernel()
78 for (size_t i = kAxis + 1; i < input_shape_.size(); i++) { in ReInit()
79 output_size *= input_shape_[i]; in ReInit()
107 const std::vector<size_t> &EmbeddingLookUpPSKernel::input_sizes() const { return input_shape_; } in input_sizes()
/third_party/mindspore/mindspore/lite/test/ut/src/runtime/kernel/arm/fp32_grad/
Dsoftmax_grad_fp32_tests.cc39 softmax_param->input_shape_[0] = 1; in InitSoftMaxParam()
40 softmax_param->input_shape_[1] = 9; in InitSoftMaxParam()
41 softmax_param->input_shape_[2] = 11; in InitSoftMaxParam()
42 softmax_param->input_shape_[3] = 12; in InitSoftMaxParam()
49 softmax_param->input_shape_[0] = n; in InitSoftMaxParam()
50 softmax_param->input_shape_[1] = c; in InitSoftMaxParam()
51 softmax_param->input_shape_[2] = h; in InitSoftMaxParam()
52 softmax_param->input_shape_[3] = w; in InitSoftMaxParam()
64 inner_size *= softmax_param->input_shape_[i]; in TEST_F()
68 …float *sum_mul = new (std::nothrow) float[inner_size * softmax_param->input_shape_[softmax_param->… in TEST_F()
[all …]
/third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/tbe/tbe_kernel_select/
Dtbe_kernel_reduce_selecter.cc35 input_shape_.clear(); in GetShapeInfo()
45 input_shape_ = AnfAlgo::GetPrevNodeOutputInferShape(cnode_ptr_, kInputIndex_0); in GetShapeInfo()
46 PadScalarShape(&input_shape_); in GetShapeInfo()
59 if (!Is4DShape(input_shape_)) { in IsReduceSupport5HD()
75 if (!Is5DShape(input_shape_)) { in IsReduceSupportNDC1HWC0()
101 if (input_shape_.size() < kReduceNZMinDim) { in IsReduceSupportFracNZ()
108 return (elem == (this->input_shape_.size() - 1) || elem == (this->input_shape_.size() - 2)); in IsReduceSupportFracNZ()
120 if (!Is4DShape(input_shape_)) { in IsFracZAndC1HWNCoC0Common()

12345