FM / ATen

/usr/local/lib64/python3.6/site-packages/torch/include/ATen/

/usr/local/lib64/python3.6/site-packages/torch/include/ATen

Name	Size	Mode	Actions
core/	-	0755	rm
cpu/	-	0755	rm
cuda/	-	0755	rm
cudnn/	-	0755	rm
detail/	-	0755	rm
hip/	-	0755	rm
native/	-	0755	rm
quantized/	-	0755	rm
AccumulateType.h	4438	0644	edit dl rm
ArrayRef.h	44	0644	edit dl rm
ATen.h	998	0644	edit dl rm
autocast_mode.h	6716	0644	edit dl rm
Backend.h	43	0644	edit dl rm
Backtrace.h	46	0644	edit dl rm
BatchedFallback.h	965	0644	edit dl rm
BatchedTensorImpl.h	5383	0644	edit dl rm
CompositeExplicitAutogradFunctions.h	1622	0644	edit dl rm
CompositeExplicitAutogradFunctions_inl.h	54075	0644	edit dl rm
CompositeImplicitAutogradFunctions.h	1622	0644	edit dl rm
CompositeImplicitAutogradFunctions_inl.h	142082	0644	edit dl rm
Config.h	734	0644	edit dl rm
Context.h	12767	0644	edit dl rm
cpp_custom_type_hack.h	5326	0644	edit dl rm
CPUApplyUtils.h	12582	0644	edit dl rm
CPUFixedAllocator.h	830	0644	edit dl rm
CPUFunctions.h	1600	0644	edit dl rm
CPUFunctions_inl.h	171924	0644	edit dl rm
CPUGeneratorImpl.h	1431	0644	edit dl rm
CUDAFunctions.h	1601	0644	edit dl rm
CUDAFunctions_inl.h	185696	0644	edit dl rm
CUDAGeneratorImpl.h	4695	0644	edit dl rm
Device.h	42	0644	edit dl rm
DeviceGuard.h	1134	0644	edit dl rm
Dimname.h	31	0644	edit dl rm
DimVector.h	46	0644	edit dl rm
Dispatch.h	52137	0644	edit dl rm
div_rtn.h	204	0644	edit dl rm
DLConvertor.h	576	0644	edit dl rm
dlpack.h	5244	0644	edit dl rm
DynamicLibrary.h	369	0644	edit dl rm
ExpandUtils.h	14506	0644	edit dl rm
Formatting.h	34	0644	edit dl rm
Functions.h	846326	0644	edit dl rm
Generator.h	46	0644	edit dl rm
InferSize.h	2143	0644	edit dl rm
InitialTensorOptions.h	445	0644	edit dl rm
Layout.h	42	0644	edit dl rm
MapAllocator.h	2999	0644	edit dl rm
MatrixRef.h	3016	0644	edit dl rm
MemoryOverlap.h	1117	0644	edit dl rm
MetaFunctions.h	1601	0644	edit dl rm
MetaFunctions_inl.h	84006	0644	edit dl rm
NamedTensor.h	35	0644	edit dl rm
NamedTensorUtils.h	5747	0644	edit dl rm
NativeFunctions.h	354651	0644	edit dl rm
NativeMetaFunctions.h	35445	0644	edit dl rm
NumericUtils.h	2787	0644	edit dl rm
OpaqueTensorImpl.h	6080	0644	edit dl rm
Operators.h	1707199	0644	edit dl rm
OpMathType.h	460	0644	edit dl rm
Parallel.h	4875	0644	edit dl rm
ParallelNative.h	2443	0644	edit dl rm
ParallelNativeTBB.h	2934	0644	edit dl rm
ParallelOpenMP.h	3049	0644	edit dl rm
PTThreadPool.h	394	0644	edit dl rm
record_function.h	24044	0644	edit dl rm
RedispatchFunctions.h	1112886	0644	edit dl rm
RegistrationDeclarations.h	545777	0644	edit dl rm
SavedTensorHooks.h	328	0644	edit dl rm
Scalar.h	44	0644	edit dl rm
ScalarOps.h	2272	0644	edit dl rm
ScalarType.h	129	0644	edit dl rm
SequenceNumber.h	373	0644	edit dl rm
SmallVector.h	47	0644	edit dl rm
SparseCsrTensorImpl.h	2045	0644	edit dl rm
SparseCsrTensorUtils.h	523	0644	edit dl rm
SparseTensorImpl.h	12417	0644	edit dl rm
SparseTensorUtils.h	4219	0644	edit dl rm
Storage.h	43	0644	edit dl rm
Tensor.h	48	0644	edit dl rm
TensorAccessor.h	51	0644	edit dl rm
TensorGeometry.h	1855	0644	edit dl rm
TensorIndexing.h	21923	0644	edit dl rm
TensorIterator.h	29962	0644	edit dl rm
TensorIteratorInternal.h	1862	0644	edit dl rm
TensorMeta.h	2917	0644	edit dl rm
TensorNames.h	2519	0644	edit dl rm
TensorOperators.h	3275	0644	edit dl rm
TensorOptions.h	49	0644	edit dl rm
TensorUtils.h	5687	0644	edit dl rm
ThreadLocalState.h	3289	0644	edit dl rm
TracerMode.h	5576	0644	edit dl rm
TypeDefault.h	680	0644	edit dl rm
Utils.h	5993	0644	edit dl rm
Version.h	340	0644	edit dl rm
VmapMode.h	952	0644	edit dl rm
VmapTransforms.h	7654	0644	edit dl rm
WrapDimUtils.h	3438	0644	edit dl rm
WrapDimUtilsMulti.h	768	0644	edit dl rm

Edit: /usr/local/lib64/python3.6/site-packages/torch/include/ATen/dlpack.h (5244B)

/*!
 *  Copyright (c) 2017 by Contributors
 * \file dlpack.h
 * \brief The common header of DLPack.
 */
#ifndef DLPACK_DLPACK_H_
#define DLPACK_DLPACK_H_

#ifdef __cplusplus
#define DLPACK_EXTERN_C extern "C"
#else
#define DLPACK_EXTERN_C
#endif

/*! \brief The current version of dlpack */
#define DLPACK_VERSION 040

/*! \brief DLPACK_DLL prefix for windows */
#ifdef _WIN32
#ifdef DLPACK_EXPORTS
#define DLPACK_DLL __declspec(dllexport)
#else
#define DLPACK_DLL __declspec(dllimport)
#endif
#else
#define DLPACK_DLL
#endif

#include <stdint.h>
#include <stddef.h>

#ifdef __cplusplus
extern "C" {
#endif
/*!
 * \brief The device type in DLDevice.
 */
typedef enum {
  /*! \brief CPU device */
  kDLCPU = 1,
  /*! \brief CUDA GPU device */
  kDLGPU = 2,
  /*!
   * \brief Pinned CUDA GPU device by cudaMallocHost
   * \note kDLCPUPinned = kDLCPU | kDLGPU
   */
  kDLCPUPinned = 3,
  /*! \brief OpenCL devices. */
  kDLOpenCL = 4,
  /*! \brief Vulkan buffer for next generation graphics. */
  kDLVulkan = 7,
  /*! \brief Metal for Apple GPU. */
  kDLMetal = 8,
  /*! \brief Verilog simulator buffer */
  kDLVPI = 9,
  /*! \brief ROCm GPUs for AMD GPUs */
  kDLROCM = 10,
  /*!
   * \brief Reserved extension device type,
   * used for quickly test extension device
   * The semantics can differ depending on the implementation.
   */
  kDLExtDev = 12,
} DLDeviceType;

/*!
 * \brief A Device for Tensor and operator.
 */
typedef struct {
  /*! \brief The device type used in the device. */
  DLDeviceType device_type;
  /*! \brief The device index */
  int device_id;
} DLDevice;

/*!
 * \brief This is an alias for DLDevice. Notice that this will be removed in the next release.
 */
typedef DLDevice DLContext;

/*!
 * \brief The type code options DLDataType.
 */
typedef enum {
  /*! \brief signed integer */
  kDLInt = 0U,
  /*! \brief unsigned integer */
  kDLUInt = 1U,
  /*! \brief IEEE floating point */
  kDLFloat = 2U,
  /*!
   * \brief Opaque handle type, reserved for testing purposes.
   * Frameworks need to agree on the handle data type for the exchange to be well-defined.
   */
  kDLOpaqueHandle = 3U,
  /*! \brief bfloat16 */
  kDLBfloat = 4U,
  /*!
   * \brief complex number
   * (C/C++/Python layout: compact struct per complex number)
   */
  kDLComplex = 5U,
} DLDataTypeCode;

/*!
 * \brief The data type the tensor can hold.
 *
 *  Examples
 *   - float: type_code = 2, bits = 32, lanes=1
 *   - float4(vectorized 4 float): type_code = 2, bits = 32, lanes=4
 *   - int8: type_code = 0, bits = 8, lanes=1
 */
typedef struct {
  /*!
   * \brief Type code of base types.
   * We keep it uint8_t instead of DLDataTypeCode for minimal memory
   * footprint, but the value should be one of DLDataTypeCode enum values.
   * */
  uint8_t code;
  /*!
   * \brief Number of bits, common choices are 8, 16, 32.
   */
  uint8_t bits;
  /*! \brief Number of lanes in the type, used for vector types. */
  uint16_t lanes;
} DLDataType;

/*!
 * \brief Plain C Tensor object, does not manage memory.
 */
typedef struct {
  /*!
   * \brief The opaque data pointer points to the allocated data. This will be
   * CUDA device pointer or cl_mem handle in OpenCL. This pointer is always
   * aligned to 256 bytes as in CUDA.
   *
   * For given DLTensor, the size of memory required to store the contents of
   * data is calculated as follows:
   *
   * \code{.c}
   * static inline size_t GetDataSize(const DLTensor* t) {
   *   size_t size = 1;
   *   for (tvm_index_t i = 0; i < t->ndim; ++i) {
   *     size *= t->shape[i];
   *   }
   *   size *= (t->dtype.bits * t->dtype.lanes + 7) / 8;
   *   return size;
   * }
   * \endcode
   */
  void* data;
  /*! \brief The device of the tensor */
  DLDevice device;
  /*! \brief Number of dimensions */
  int ndim;
  /*! \brief The data type of the pointer*/
  DLDataType dtype;
  /*! \brief The shape of the tensor */
  int64_t* shape;
  /*!
   * \brief strides of the tensor (in number of elements, not bytes)
   *  can be NULL, indicating tensor is compact and row-majored.
   */
  int64_t* strides;
  /*! \brief The offset in bytes to the beginning pointer to data */
  uint64_t byte_offset;
} DLTensor;

/*!
 * \brief C Tensor object, manage memory of DLTensor. This data structure is
 *  intended to facilitate the borrowing of DLTensor by another framework. It is
 *  not meant to transfer the tensor. When the borrowing framework doesn't need
 *  the tensor, it should call the deleter to notify the host that the resource
 *  is no longer needed.
 */
typedef struct DLManagedTensor {
  /*! \brief DLTensor which is being memory managed */
  DLTensor dl_tensor;
  /*! \brief the context of the original host framework of DLManagedTensor in
   *   which DLManagedTensor is used in the framework. It can also be NULL.
   */
  void * manager_ctx;
  /*! \brief Destructor signature void (*)(void*) - this should be called
   *   to destruct manager_ctx which holds the DLManagedTensor. It can be NULL
   *   if there is no way for the caller to provide a reasonable destructor.
   *   The destructors deletes the argument self as well.
   */
  void (*deleter)(struct DLManagedTensor * self);
} DLManagedTensor;
#ifdef __cplusplus
}  // DLPACK_EXTERN_C
#endif
#endif  // DLPACK_DLPACK_H_

cmd: