/aosp_15_r20/external/tensorflow/tensorflow/lite/delegates/gpu/common/ |
H A D | gpu_model.cc | 266 GpuModel* gpu_model) { in ConvertOperations() argument 276 for (const auto& input : gpu_model->input_ids_and_refs) { in ConvertOperations() 295 gpu_model->const_tensors[outputs[0]->id] = in ConvertOperations() 297 gpu_model->const_tensors[outputs[0]->id].UploadData(attr.tensor); in ConvertOperations() 343 gpu_model->const_tensors[global_id] = in ConvertOperations() 377 gpu_model->nodes.push_back(std::move(gpu_node)); in ConvertOperations() 385 GpuModel* gpu_model) { in MergeElementwiseNodes() argument 386 auto& nodes = gpu_model->nodes; in MergeElementwiseNodes() 637 absl::Status MergeNodes(const GpuInfo& gpu_info, GpuModel* gpu_model) { in MergeNodes() argument 639 for (const auto& input : gpu_model->input_ids_and_refs) { in MergeNodes() [all …]
|
H A D | gpu_model_test_util.cc | 79 GpuModel gpu_model; in TestLinkingConvolutionAndCosOp() local 81 GraphToGpuModel(graph, create_info, env->GetGpuInfo(), &gpu_model)); in TestLinkingConvolutionAndCosOp() 83 if (gpu_model.nodes.size() != 1) { in TestLinkingConvolutionAndCosOp() 96 env->ExecuteGpuModel({src_tensor}, {&dst_tensor_v1}, &gpu_model)); in TestLinkingConvolutionAndCosOp() 191 GpuModel gpu_model; in TestLinkingConvolution2InputMul2InputMul() local 193 GraphToGpuModel(graph, create_info, env->GetGpuInfo(), &gpu_model)); in TestLinkingConvolution2InputMul2InputMul() 195 if (gpu_model.nodes.size() != 1) { in TestLinkingConvolution2InputMul2InputMul() 221 {&dst_tensor_v1}, &gpu_model)); in TestLinkingConvolution2InputMul2InputMul() 330 GpuModel gpu_model; in TestLinkingConvolution2InputBroadcastMul2InputMul() local 332 GraphToGpuModel(graph, create_info, env->GetGpuInfo(), &gpu_model)); in TestLinkingConvolution2InputBroadcastMul2InputMul() [all …]
|
H A D | gpu_model.h | 109 const GpuInfo& gpu_info, GpuModel* gpu_model); 112 const GpuModel& gpu_model, flatbuffers::FlatBufferBuilder* builder); 114 absl::Status Decode(const data::GpuModel* fb_gpu_model, GpuModel* gpu_model);
|
H A D | BUILD | 51 name = "gpu_model", 52 srcs = ["gpu_model.cc"], 53 hdrs = ["gpu_model.h"], 75 srcs = ["gpu_model.fbs"], 91 ":gpu_model",
|
/aosp_15_r20/external/tensorflow/tensorflow/lite/delegates/gpu/cl/ |
H A D | inference_context.cc | 86 TensorType GetTensorType(const GpuModel& gpu_model, in GetTensorType() argument 90 for (int i = 0; i < gpu_model.variable_ids_and_refs.size(); ++i) { in GetTensorType() 91 if (gpu_model.variable_ids_and_refs[i].first == id) { in GetTensorType() 104 } else if (gpu_model.const_tensors.find(id) != in GetTensorType() 105 gpu_model.const_tensors.end()) { in GetTensorType() 140 const GpuModel& gpu_model, const CreateGpuModelInfo* create_info, in GetBufferAsignment() argument 149 gpu_model, in GetBufferAsignment() 150 [&gpu_model, &gpu_info, &create_info](ValueId id) { in GetBufferAsignment() 151 return GetTensorType(gpu_model, create_info, gpu_info, id) == in GetBufferAsignment() 154 gpu_model.tensors.at(id).GetStorageType()); in GetBufferAsignment() [all …]
|
H A D | inference_context.h | 73 const CreateGpuModelInfo& create_info, GpuModel* gpu_model, 119 void InitFromGpuModel(GpuModel* gpu_model); 121 absl::Status AllocateMemory(const GpuModel& gpu_model, 126 absl::Status AllocateConstTensors(const GpuModel& gpu_model, 129 absl::Status AllocateVariableTensors(const GpuModel& gpu_model, 132 absl::Status AllocateBufferBasedTensors(const GpuModel& gpu_model, 138 const GpuModel& gpu_model, const GpuInfo& gpu_info, 207 absl::Status GetTotalBufferSizeForTensors(const GpuModel& gpu_model,
|
H A D | serialization_generated.h | 109 const tflite::gpu::data::GpuModel *gpu_model() const { in gpu_model() function 127 verifier.VerifyTable(gpu_model()) && in Verify() 146 void add_gpu_model(flatbuffers::Offset<tflite::gpu::data::GpuModel> gpu_model) { in add_gpu_model() 147 fbb_.AddOffset(InferenceContext::VT_GPU_MODEL, gpu_model); in add_gpu_model() 174 flatbuffers::Offset<tflite::gpu::data::GpuModel> gpu_model = 0, 184 builder_.add_gpu_model(gpu_model); 190 flatbuffers::Offset<tflite::gpu::data::GpuModel> gpu_model = 0, 201 gpu_model,
|
H A D | serialization.fbs | 15 include "tensorflow/lite/delegates/gpu/common/gpu_model.fbs"; 26 gpu_model:tflite.gpu.data.GpuModel;
|
H A D | BUILD | 403 "//tensorflow/lite/delegates/gpu/common:gpu_model",
|
/aosp_15_r20/external/tensorflow/tensorflow/lite/delegates/gpu/metal/ |
H A D | inference_context.cc | 122 void InferenceContext::CopyFromGpuModel(GpuModel* gpu_model) { in CopyFromGpuModel() argument 123 for (const auto& input : gpu_model->input_ids_and_refs) { in CopyFromGpuModel() 126 for (const auto& output : gpu_model->output_ids_and_refs) { in CopyFromGpuModel() 129 nodes_.resize(gpu_model->nodes.size()); in CopyFromGpuModel() 130 for (int i = 0; i < gpu_model->nodes.size(); ++i) { in CopyFromGpuModel() 131 nodes_[i].task.Init(std::move(gpu_model->nodes[i].gpu_operation)); in CopyFromGpuModel() 132 nodes_[i].inputs = gpu_model->nodes[i].inputs; in CopyFromGpuModel() 133 nodes_[i].outputs = gpu_model->nodes[i].outputs; in CopyFromGpuModel() 134 nodes_[i].name = gpu_model->nodes[i].name; in CopyFromGpuModel() 136 const_tensors_descs_ = std::move(gpu_model->const_tensors); in CopyFromGpuModel() [all …]
|
H A D | inference_context.fbs | 15 include "tensorflow/lite/delegates/gpu/common/gpu_model.fbs"; 27 gpu_model:tflite.gpu.data.GpuModel;
|
H A D | inference_context.h | 152 void CopyFromGpuModel(GpuModel* gpu_model);
|
H A D | BUILD | 153 "//tensorflow/lite/delegates/gpu/common:gpu_model",
|
/aosp_15_r20/external/pytorch/test/distributed/algorithms/ddp_comm_hooks/ |
H A D | test_ddp_hooks.py | 95 gpu_model = DistributedDataParallel( 104 comm_hook_type=hook_type, model=gpu_model, state=process_group 107 return self._run_and_get_grads(gpu_model) 224 gpu_model = DistributedDataParallel( 229 gpu_model.register_comm_hook(state=flags, hook=hook) 231 gpu_model(input).sum().backward()
|
/aosp_15_r20/external/tensorflow/tensorflow/lite/delegates/gpu/common/task/ |
H A D | testing_util.cc | 131 const std::vector<TensorFloat32*>& dst_cpu, GpuModel* gpu_model) { in ExecuteGpuModel() argument 133 for (int k = 0; k < gpu_model->nodes.size(); ++k) { in ExecuteGpuModel() 134 auto& gpu_node = gpu_model->nodes[k]; in ExecuteGpuModel() 138 src_cpu_descs[i] = gpu_model->tensors[gpu_node.inputs[i]]; in ExecuteGpuModel() 146 dst_cpu_descs[i] = gpu_model->tensors[gpu_node.outputs[i]]; in ExecuteGpuModel() 152 std::move(gpu_model->nodes[k].gpu_operation))); in ExecuteGpuModel()
|
H A D | testing_util.h | 53 GpuModel* gpu_model);
|
H A D | BUILD | 152 "//tensorflow/lite/delegates/gpu/common:gpu_model",
|
/aosp_15_r20/external/pytorch/test/distributed/ |
H A D | test_c10d_ucc.py | 465 gpu_model = DistributedDataParallel( 473 run_and_verify_grad(gpu_model) 546 gpu_model = DistributedDataParallel( 552 run_and_verify_grad(gpu_model) 818 gpu_model = DistributedDataParallel( 827 gpu_model.register_comm_hook(state, hook) 829 return gpu_model 841 gpu_model = self._gpu_model_with_ddp_comm_hook(process_group, self._simple_hook) 845 self._run_and_verify_hook(gpu_model, 8, 2 * torch.ones(2, 2))
|
H A D | test_c10d_nccl.py | 1752 gpu_model = DistributedDataParallel( 1762 gpu_model.register_comm_hook(state, hook) 1764 return gpu_model 1776 gpu_model = self._gpu_model_with_ddp_comm_hook(process_group, self._simple_hook) 1780 self._run_and_verify_hook(gpu_model, 8, 2 * torch.ones(2, 2)) 1804 gpu_model = self._gpu_model_with_ddp_comm_hook( 1809 self._run_and_verify_hook(gpu_model, 8, 0.25 * torch.ones(2, 2)) 1832 gpu_model = self._gpu_model_with_ddp_comm_hook( 1837 self._run_and_verify_hook(gpu_model, 8, 0.25 * torch.ones(2, 2)) 1853 gpu_model = self._gpu_model_with_ddp_comm_hook( [all …]
|
H A D | test_c10d_gloo.py | 1606 gpu_model = DistributedDataParallel( 1614 run_and_verify_grad(gpu_model) 1679 gpu_model = DistributedDataParallel( 1685 run_and_verify_grad(gpu_model) 1990 gpu_model = DistributedDataParallel( 1999 gpu_model.register_comm_hook(state, hook) 2001 return gpu_model 2013 gpu_model = self._gpu_model_with_ddp_comm_hook(process_group, self._simple_hook) 2017 self._run_and_verify_hook(gpu_model, 8, 2 * torch.ones(2, 2))
|
H A D | test_c10d_common.py | 832 gpu_model = DistributedDataParallel( 841 gpu_model.register_comm_hook(state, hook) 843 return gpu_model 849 gpu_model = DistributedDataParallel( 858 gpu_model._register_builtin_comm_hook(hook) 860 return gpu_model
|
/aosp_15_r20/external/tensorflow/tensorflow/core/profiler/convert/ |
H A D | xplane_to_op_stats.cc | 107 absl::string_view gpu_model = in SetRunEnvironment() local 109 if (!gpu_model.empty()) { in SetRunEnvironment() 110 env->set_device_type(std::string(gpu_model)); in SetRunEnvironment()
|
/aosp_15_r20/external/pytorch/docs/source/ |
H A D | jit.rst | 564 cpu_model = gpu_model.cpu() 569 traced_gpu = torch.jit.trace(gpu_model, sample_input_gpu)
|
/aosp_15_r20/external/cpuinfo/test/dmesg/ |
H A D | moto-g-gen5.log | 1427 …708.153:60): avc: denied { read } for uid=2000 pid=2682 comm="sh" name="gpu_model" dev="sysfs" ino…
|