File tree

• .circleci
  • cimodel/data
    • binary_build_data.py
    • dimensions.py
    • simple
      • docker_definitions.py
  • config.yml
  • docker
    • build.sh
    • centos-rocm
      • Dockerfile
    • common
      • install_base.sh
      • install_conda.sh
      • install_travis_python.sh
    • ubuntu-cuda
      • Dockerfile
    • ubuntu-rocm
      • Dockerfile
    • ubuntu
      • Dockerfile
  • scripts
    • binary_linux_test.sh
  • verbatim-sources/job-specs
    • pytorch-job-specs.yml
• .github/workflows
  • lint.yml
• .gitignore
• .jenkins
  • caffe2
    • build.sh
    • common.sh
    • test.sh
  • pytorch
    • build.sh
    • codegen-test.sh
    • common.sh
    • macos-build.sh
    • multigpu-test.sh
    • test.sh
    • win-test.sh
• .nojekyll
• CMakeLists.txt
• CONTRIBUTING.md
• README.md
• aten/src
  • ATen
    • BatchedTensorImpl.cpp
    • BatchingRegistrations.cpp
    • CMakeLists.txt
    • Dispatch.h
    • LegacyTHFunctionsCPU.cpp
    • LegacyTHFunctionsCPU.h
    • NumericUtils.h
    • TensorIndexing.h
    • benchmarks
      • quantize_per_channel.cpp
    • core
      • Dimname.h
      • Generator.h
      • NamedRegistrations.cpp
      • NamedTensor.h
      • aten_interned_strings.h
      • boxing
        • KernelFunction_test.cpp
        • impl
          • boxing.h
          • make_boxed_from_unboxed_functor.h
      • builtin_function.h
      • dispatch
        • Dispatcher.h
        • OperatorEntry.cpp
        • OperatorEntry.h
      • function.h
      • interned_strings.h
      • ivalue.cpp
      • ivalue.h
      • op_registration
        • op_registration_test.cpp
    • cpu/vec256
      • vec256_bfloat16.h
      • vec256_complex_double.h
      • vec256_complex_float.h
      • vec256_double.h
      • vec256_float.h
      • vec256_float_neon.h
    • cuda
      • CUDABlas.cpp
      • CUDABlas.h
      • CUDASolver.cpp
      • CUDASolver.h
    • native
      • BinaryOps.cpp
      • BinaryOps.h
      • Copy.cpp
      • Embedding.cpp
      • EmbeddingBag.cpp
      • ForeachOpsKernels.cpp
      • ForeachUtils.h
      • GridSampler.cpp
      • GridSampler.h
      • Linear.cpp
      • LinearAlgebra.cpp
      • ReduceOps.cpp
      • Resize.h
      • ResizeCommon.h
      • SharedReduceOps.h
      • TensorAdvancedIndexing.cpp
      • TensorFactories.cpp
      • TensorIterator.cpp
      • TensorShape.cpp
      • TensorTransformations.cpp
      • UnaryOps.cpp
      • VariableMethodStubs.cpp
      • cpu
        • BinaryOpsKernel.cpp
        • DistributionTemplates.h
        • GridSamplerKernel.cpp
        • MaxPooling.cpp
        • ReduceOpsKernel.cpp
        • TensorCompareKernel.cpp
      • cuda
        • Activation.cu
        • AdaptiveAveragePooling.cu
        • AdaptiveAveragePooling3d.cu
        • AdaptiveMaxPooling2d.cu
        • AdaptiveMaxPooling3d.cu
        • AmpKernels.cu
        • AveragePool2d.cu
        • BatchLinearAlgebra.cu
        • BatchLinearAlgebraLib.cu
        • BinaryGeometricKernels.cu
        • BinaryMiscOpsKernels.cu
        • Bucketization.cu
        • CopysignKernel.cu
        • DilatedMaxPool2d.cu
        • DilatedMaxPool3d.cu
        • DistanceKernel.cu
        • Dropout.cu
        • Embedding.cu
        • EmbeddingBackwardKernel.cu
        • EmbeddingBag.cu
        • ForeachBinaryOpList.cu
        • ForeachBinaryOpScalar.cu
        • ForeachBinaryOpScalarList.cu
        • ForeachFunctors.cuh
        • ForeachPointwiseOp.cu
        • ForeachUnaryOp.cu
        • FractionalMaxPool2d.cu
        • FractionalMaxPool3d.cu
        • FunctionOfAMatrixUtilsKernel.cu
        • GcdLcmKernel.cu
        • GridSampler.cuh
        • GridSampler.cu
        • IGammaKernel.cu
        • IndexKernel.cu
        • Indexing.cu
        • LinearAlgebra.cu
        • LogAddExpKernel.cu
        • Loops.cuh
        • LossCTC.cu
        • Math.cuh
        • MaxUnpooling.cu
        • MultiTensorApply.cuh
        • ReduceLogicKernel.cu
        • ReduceMomentKernel.cu
        • ReduceSumProdKernel.cu
        • Resize.cuh
        • Shape.cu
        • Sorting.cu
        • StepKernel.cu
        • TensorFactories.cu
        • TensorTransformations.cu
        • TriangularOps.cu
        • UnaryGeometricKernels.cu
        • UnfoldBackwardKernel.cu
        • UpSampleBicubic2d.cu
        • UpSampleBilinear2d.cu
        • UpSampleLinear1d.cu
        • UpSampleNearest1d.cu
        • UpSampleNearest2d.cu
        • UpSampleNearest3d.cu
        • UpSampleTrilinear3d.cu
        • group_norm_kernel.cu
        • im2col.cuh
        • layer_norm_kernel.cu
        • vol2col.cuh
      • metal
        • MetalAten.mm
        • MetalPrepackOpRegister.cpp
        • mpscnn/tests
          • MPSCNNTests.mm
      • native_functions.yaml
      • quantized
        • affine_quantizer.cpp
        • cpu
          • embedding_packed_params.h
          • fbgemm_utils.h
          • kernels
            • QuantizedOpKernels.cpp
          • qembeddingbag.cpp
        • library.cpp
      • vulkan
        • Vulkan.cpp
        • VulkanAten.cpp
        • VulkanOps.cpp
        • VulkanRegisterOpContextClass.cpp
        • api
          • Command.cpp
          • Resource.cpp
          • Utils.h
        • glsl
          • KO4C4HW_to_image.glsl
          • adaptive_avg_pool2d.glsl
          • add.glsl
          • add_.glsl
          • add_scalar.glsl
          • add_scalar_.glsl
          • addmm.glsl
          • avg_pool2d.glsl
          • clamp.glsl
          • clamp_.glsl
          • conv2d.glsl
          • conv2d_dw.glsl
          • conv2d_dw_clamp.glsl
          • conv2d_nogroup_clamp.glsl
          • conv2d_pw.glsl
          • image_to_nchw.glsl
          • max_pool2d.glsl
          • mean.glsl
          • mean2d.glsl
          • mm.glsl
          • mul_scalar.glsl
          • mul_scalar_.glsl
          • nchw_to_image.glsl
          • upsampleNearest2d.glsl
          • upsample_nearest2d.glsl
        • ops
          • Add.cpp
          • Clamp.cpp
          • Common.h
          • Convolution.cpp
          • Factory.cpp
          • Mean.cpp
          • Mm.cpp
          • Mul.cpp
          • Persistent.cpp
          • Persistent.h
          • Pool.cpp
          • Shape.cpp
          • Tensor.cpp
          • Tensor.h
          • Upsample.cpp
      • xnnpack
        • OpContext.cpp
        • OpContext.h
    • nnapi
      • CMakeLists.txt
      • NeuralNetworks.h
      • codegen.py
      • nnapi_bind.cpp
      • nnapi_model_loader.cpp
      • nnapi_model_loader.h
      • nnapi_wrapper.cpp
      • nnapi_wrapper.h
    • templates
      • TensorBody.h
    • test
      • CMakeLists.txt
      • cuda_atomic_ops_test.cu
      • ivalue_test.cpp
      • mobile_memory_cleanup.cpp
      • quantized_test.cpp
      • scalar_test.cpp
      • vec256_test_all_types.h
      • vmap_test.cpp
      • vulkan_api_test.cpp
      • vulkan_test.cpp
  • TH/generic
    • THTensorMath.h
    • THTensorMoreMath.cpp
• benchmarks
  • operator_benchmark
    • benchmark_caffe2.py
    • c2
      • batch_box_cox_test.py
    • pt
      • qembedding_bag_lookups_test.py
      • tensor_to_test.py
  • static_runtime
    • deep_wide_pt.cc
    • test_static_runtime.cc
• c10
  • core
    • DispatchKey.cpp
    • DispatchKey.h
    • DispatchKeySet.cpp
    • DispatchKeySet.h
    • Scalar.cpp
    • Scalar.h
  • cuda
    • CUDAMathCompat.h
    • CUDAStream.cpp
    • CUDAStream.h
  • test/util
    • Metaprogramming_test.cpp
    • optional_test.cpp
  • util
    • C++17.h
    • Half.h
    • Metaprogramming.h
    • Optional.cpp
    • Optional.h
    • intrusive_ptr.h
• caffe2
  • contrib/nccl
    • cuda_nccl_gpu.cc
  • operators
    • box_with_nms_limit_op.cc
    • box_with_nms_limit_op.h
    • lengths_reducer_fused_nbit_rowwise_ops.cc
    • lengths_reducer_fused_nbit_rowwise_ops.h
    • self_binning_histogram_op.cc
    • self_binning_histogram_op.h
  • perfkernels
    • embedding_lookup_fused_8bit_rowwise_idx_avx2.cc
    • embedding_lookup_idx.cc
    • embedding_lookup_idx.h
    • embedding_lookup_idx_avx2.cc
    • fused_8bit_rowwise_embedding_lookup_idx.cc
    • fused_8bit_rowwise_embedding_lookup_idx.h
    • hp_emblookup_codegen.py
  • python
    • brew.py
    • core.py
    • core_test.py
    • helpers
      • algebra.py
    • hypothesis_test.py
    • operator_test
      • self_binning_histogram_test.py
      • torch_integration_test.py
    • schema.py
    • schema_test.py
  • utils/threadpool
    • ThreadPool.cc
• cmake
  • Dependencies.cmake
  • Modules_CUDA_fix/upstream/FindCUDA
    • select_compute_arch.cmake
• codecov.yml
• docker/caffe2/jenkins
  • centos-rocm
    • Dockerfile
  • ubuntu-rocm
    • Dockerfile
• docs/source
  • benchmark_utils.rst
  • index.rst
  • jit.rst
  • notes
    • cuda.rst
  • tensors.rst
  • torch.rst
• mypy.ini
• scripts
  • onnx
    • test.sh
  • release_notes
    • categorize.py
    • commitlist.py
    • common.py
    • requirements.txt
    • test_release_notes.py
• setup.py
• test
  • backward_compatibility
    • check_backward_compatibility.py
  • cpp_extensions
    • cpp_c10d_extension.cpp
    • cpp_c10d_extension.hpp
  • cpp
    • api
      • autograd.cpp
      • modules.cpp
    • jit
      • test_autodiff.cpp
      • test_graph_executor.cpp
      • test_interpreter.cpp
      • test_interpreter_async.pt
      • test_lite_interpreter.cpp
      • test_subgraph_utils.cpp
    • rpc
      • e2e_test_base.h
    • tensorexpr
      • test_boundsinference.cpp
      • test_kernel.cpp
      • test_llvm.cpp
      • test_memdependency.cpp
      • test_reductions.cpp
      • tests.h
  • distributed
    • test_c10d.py
  • jit
    • test_list_dict.py
    • test_recursive_script.py
    • test_save_load.py
    • test_torchbind.py
  • mobile/custom_build
    • CMakeLists.txt
  • onnx
    • expect
      • TestOperators.test_std.expect
    • test_pytorch_onnx_onnxruntime.py
  • quantization
    • test_quantize.py
    • test_quantize_fx.py
    • test_quantized_op.py
    • test_workflow_module.py
  • run_test.py
  • test_autograd.py
  • test_cuda.py
  • test_foreach.py
  • test_fx.py
  • test_fx_experimental.py
  • test_jit.py
  • test_linalg.py
  • test_nn.py
  • test_overrides.py
  • test_sparse.py
  • test_static_runtime.py
  • test_tensor_creation_ops.py
  • test_tensorexpr.py
  • test_torch.py
  • test_unary_ufuncs.py
  • test_vmap.py
• third_party
  • fbgemm
  • nccl
    • nccl
• tools
  • autograd
    • derivatives.yaml
    • gen_autograd.py
    • gen_python_functions.py
    • gen_trace_type.py
    • gen_variable_type.py
    • load_derivatives.py
    • templates
      • python_torch_functions.cpp
      • python_variable_methods.cpp
  • build_variables.bzl
  • code_analyzer
    • gen_op_registration_allowlist.py
  • codegen
    • api
      • cpp.py
      • python.py
    • gen.py
    • model.py
  • pyi
    • gen_pyi.py
  • setup_helpers
    • generate_code.py
• torch
  • _C
    • __init__.pyi.in
    • _autograd.pyi
    • _distributed_autograd.pyi
    • _distributed_c10d.pyi
    • _distributed_rpc.pyi
  • _jit_internal.py
  • _lowrank.py
  • _tensor_docs.py
  • _torch_docs.py
  • autograd
    • __init__.py
    • profiler.py
  • backends/_nnapi
    • __init__.py
    • prepare.py
    • serializer.py
  • csrc
    • api/include/torch/nn/functional
      • vision.h
    • autograd
      • FunctionsManual.cpp
      • FunctionsManual.h
      • TraceTypeManual.cpp
      • VariableTypeManual.cpp
      • anomaly_mode.cpp
      • anomaly_mode.h
      • autograd.cpp
      • autograd.h
      • engine.cpp
      • engine.h
      • functions
        • tensor.cpp
      • python_engine.cpp
      • python_engine.h
      • saved_variable.cpp
    • distributed
      • autograd
        • engine
          • dist_engine.cpp
        • init.cpp
        • rpc_messages
          • rref_backward_req.cpp
          • rref_backward_req.h
          • rref_backward_resp.cpp
          • rref_backward_resp.h
      • c10d
        • init.cpp
        • python_comm_hook.cpp
        • python_comm_hook.h
      • rpc
        • init.cpp
        • message.cpp
        • message.h
        • process_group_agent.cpp
        • process_group_agent.h
        • py_rref.cpp
        • py_rref.h
        • request_callback_impl.cpp
        • request_callback_impl.h
        • request_callback_no_python.cpp
        • request_callback_no_python.h
        • tensorpipe_agent.cpp
        • tensorpipe_agent.h
        • utils.cpp
    • jit
      • api
        • function_impl.cpp
        • function_impl.h
      • frontend
        • ir_emitter.cpp
      • ir
        • ir.cpp
        • scope.cpp
        • scope.h
      • mobile
        • import.cpp
        • import.h
        • module.h
      • passes
        • freeze_module.cpp
        • freeze_module.h
        • metal_rewrite.cpp
        • onnx
          • fixup_onnx_controlflow.cpp
          • preprocess_for_onnx.cpp
          • remove_inplace_ops_for_onnx.cpp
        • quantization
          • insert_quant_dequant.cpp
        • tensorexpr_fuser.cpp
        • utils
          • subgraph_utils.cpp
          • subgraph_utils.h
      • python
        • pybind_utils.h
        • python_sugared_value.cpp
        • script_init.cpp
      • runtime
        • graph_executor.cpp
        • graph_executor.h
        • graph_executor_impl.h
        • interpreter.cpp
        • interpreter.h
        • profiling_graph_executor_impl.cpp
        • profiling_graph_executor_impl.h
        • static
          • impl.cpp
          • impl.h
          • init.cpp
          • ops.cpp
          • ops.h
      • serialization
        • python_print.cpp
      • tensorexpr
        • analysis.h
        • block_codegen.h
        • bounds_inference.cpp
        • bounds_inference.h
        • bounds_overlap.cpp
        • bounds_overlap.h
        • codegen.cpp
        • codegen.h
        • cuda_codegen.cpp
        • cuda_codegen.h
        • cuda_half_support.h
        • kernel.cpp
        • kernel.h
        • llvm_codegen.cpp
        • llvm_codegen.h
        • llvm_jit.cpp
        • llvm_jit.h
        • loopnest.cpp
        • loopnest.h
        • mem_dependency_checker.cpp
        • mem_dependency_checker.h
        • reduction.h
        • types.h
    • utils
      • pybind.h
      • python_arg_parser.cpp
      • python_arg_parser.h
      • tensor_new.cpp
      • tensor_numpy.cpp
      • tensor_numpy.h
  • distributed
    • CONTRIBUTING.md
    • __init__.py
    • algorithms
      • __init__.py
      • ddp_comm_hooks
        • __init__.py
        • default_hooks.py
        • quantization_hooks.py
    • autograd
      • __init__.py
    • benchmarks
      • README.md
      • benchmark_ddp_rpc.py
    • distributed_c10d.py
    • rendezvous.py
    • rpc
      • __init__.py
      • api.py
      • backend_registry.py
      • constants.py
      • functions.py
      • internal.py
      • options.py
  • functional.py
  • fx
    • __init__.py
    • experimental
      • GraphManipulation.py
      • Partitioner.py
      • partitioner_utils.py
      • rewriter.py
    • graph.py
    • graph_module.py
    • node.py
    • proxy.py
    • symbolic_trace.py
  • jit
    • _script.py
    • _trace.py
    • annotations.py
  • lib/c10d
    • CMakeLists.txt
    • PrefixStore.cpp
    • PrefixStore.hpp
    • ProcessGroup.cpp
    • ProcessGroup.hpp
    • ProcessGroupGloo.cpp
    • ProcessGroupGloo.hpp
    • ProcessGroupMPI.cpp
    • ProcessGroupMPI.hpp
    • ProcessGroupNCCL.cpp
    • ProcessGroupNCCL.hpp
    • ProcessGroupRoundRobin.cpp
    • ProcessGroupRoundRobin.hpp
    • Store.hpp
    • comm.cpp
    • comm.hpp
    • default_comm_hooks.cpp
    • default_comm_hooks.hpp
    • example
      • allreduce.cpp
    • frontend.cpp
    • frontend.hpp
    • reducer.cpp
    • reducer.hpp
    • test
      • FileStoreTest.cpp
      • HashStoreTest.cpp
      • ProcessGroupGlooAsyncTest.cpp
      • ProcessGroupGlooTest.cpp
      • ProcessGroupMPITest.cpp
      • ProcessGroupNCCLErrorsTest.cpp
      • ProcessGroupNCCLTest.cpp
      • TCPStoreTest.cpp
  • nn
    • functional.pyi.in
    • functional.py
    • intrinsic
      • __init__.py
      • modules
        • __init__.py
        • fused.py
      • qat
        • __init__.py
        • modules
          • __init__.py
          • conv_fused.py
          • linear_relu.py
    • modules
      • batchnorm.py
      • conv.py
      • instancenorm.py
      • linear.py
      • sparse.py
    • parallel
      • distributed.py
    • parameter.pyi
    • qat/modules
      • conv.py
      • linear.py
    • quantized
      • dynamic/modules
        • linear.py
      • modules
        • conv.py
    • utils
      • rnn.py
  • onnx
    • symbolic_helper.py
    • symbolic_opset11.py
    • symbolic_opset9.py
    • utils.py
  • overrides.py
  • quantization
    • fake_quantize.py
    • fuser_method_mappings.py
    • fx
      • fuse.py
      • fusion_patterns.py
      • quantization_patterns.py
      • quantize.py
      • utils.py
    • observer.py
    • quantization_mappings.py
    • quantize.py
    • quantize_fx.py
    • utils.py
  • storage.py
  • tensor.py
  • testing
    • __init__.py
    • _internal
      • common_device_type.py
      • common_methods_invocations.py
      • common_utils.py
      • distributed
        • distributed_test.py
        • rpc
          • rpc_test.py
  • utils
    • benchmark/utils
      • timer.py
    • collect_env.py
    • cpp_extension.py
    • show_pickle.py