Skip to content

P
Paddle

Crayon鑫 / Paddle
与 Fork 源项目一致

Fork自 PaddlePaddle / Paddle

通知 1
Star 1
Fork 0
P
Paddle
“fc374821ddb9d40daaaf443c3d78ac2d3643ce03”上不存在“paddle/fluid/framework/ir/generate_pass.h”
未验证 提交 fcd32950 编写于 作者: Z zyfncg 提交者： GitHub
浏览文件
操作

Rename yaml (#43470) 

* rename yaml file

* fix merge conflict

* fix infrt
上级 346efe96
显示空白变更内容
内联 并排
Showing with 4724 addition and 4712 deletion
paddle/fluid/eager/auto_code_generator/final_state_generator/CMakeLists.txt
浏览文件 @ fcd32950
set(api_yaml_path set(api_yaml_path
"${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/api.yaml,${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/new_api.yaml,${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/sparse_api.yaml" "${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/api.yaml,${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/legacy_api.yaml,${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/sparse_api.yaml"
) )
set(backward_yaml_path set(backward_yaml_path
"${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/backward.yaml,${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/new_backward.yaml,${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/sparse_bw_api.yaml" "${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/backward.yaml,${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/legacy_backward.yaml,${PADDLE_SOURCE_DIR}/python/paddle/utils/code_gen/sparse_bw_api.yaml"
) )
set(tmp_forwards_cc_path set(tmp_forwards_cc_path
"${PADDLE_SOURCE_DIR}/paddle/fluid/eager/api/generated/eager_generated/forwards/tmp_dygraph_functions.cc" "${PADDLE_SOURCE_DIR}/paddle/fluid/eager/api/generated/eager_generated/forwards/tmp_dygraph_functions.cc"
......
paddle/phi/api/lib/CMakeLists.txt
浏览文件 @ fcd32950
...@@ -22,8 +22,8 @@ set(api_gen_base ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/api_base.py) ...@@ -22,8 +22,8 @@ set(api_gen_base ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/api_base.py)
# forward api file # forward api file
set(api_gen_file ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/api_gen.py) set(api_gen_file ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/api_gen.py)
set(api_yaml_file ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/api.yaml) set(api_yaml_file ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/api.yaml)
set(new_api_yaml_file set(legacy_api_yaml_file
${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/new_api.yaml) ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/legacy_api.yaml)
set(api_header_file ${CMAKE_SOURCE_DIR}/paddle/phi/api/include/api.h) set(api_header_file ${CMAKE_SOURCE_DIR}/paddle/phi/api/include/api.h)
set(api_source_file ${CMAKE_SOURCE_DIR}/paddle/phi/api/lib/api.cc) set(api_source_file ${CMAKE_SOURCE_DIR}/paddle/phi/api/lib/api.cc)
set(api_header_file_tmp ${api_header_file}.tmp) set(api_header_file_tmp ${api_header_file}.tmp)
...@@ -34,8 +34,8 @@ set(bw_api_gen_file ...@@ -34,8 +34,8 @@ set(bw_api_gen_file
${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/backward_api_gen.py) ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/backward_api_gen.py)
set(bw_api_yaml_file set(bw_api_yaml_file
${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/backward.yaml) ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/backward.yaml)
set(new_bw_api_yaml_file set(legacy_bw_api_yaml_file
${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/new_backward.yaml) ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen/legacy_backward.yaml)
set(bw_api_header_file set(bw_api_header_file
${CMAKE_SOURCE_DIR}/paddle/phi/api/backward/backward_api.h) ${CMAKE_SOURCE_DIR}/paddle/phi/api/backward/backward_api.h)
set(bw_api_source_file ${CMAKE_SOURCE_DIR}/paddle/phi/api/lib/backward_api.cc) set(bw_api_source_file ${CMAKE_SOURCE_DIR}/paddle/phi/api/lib/backward_api.cc)
...@@ -111,21 +111,21 @@ set(generated_argument_mapping_path ...@@ -111,21 +111,21 @@ set(generated_argument_mapping_path
message( message(
"parse api yamls: "parse api yamls:
- ${api_yaml_file} - ${api_yaml_file}
- ${new_api_yaml_file} - ${legacy_api_yaml_file}
- ${bw_api_yaml_file} - ${bw_api_yaml_file}
- ${new_bw_api_yaml_file}") - ${legacy_bw_api_yaml_file}")
execute_process( execute_process(
WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen
COMMAND ${CMAKE_COMMAND} -E make_directory ${parsed_api_dir} COMMAND ${CMAKE_COMMAND} -E make_directory ${parsed_api_dir}
COMMAND ${PYTHON_EXECUTABLE} parse_api.py --api_yaml_path ./api.yaml COMMAND ${PYTHON_EXECUTABLE} parse_api.py --api_yaml_path ./api.yaml
--output_path ./parsed_apis/api.parsed.yaml --output_path ./parsed_apis/api.parsed.yaml
COMMAND ${PYTHON_EXECUTABLE} parse_api.py --api_yaml_path ./new_api.yaml COMMAND ${PYTHON_EXECUTABLE} parse_api.py --api_yaml_path ./legacy_api.yaml
--output_path ./parsed_apis/new_api.parsed.yaml --output_path ./parsed_apis/legacy_api.parsed.yaml
COMMAND ${PYTHON_EXECUTABLE} parse_api.py --api_yaml_path ./backward.yaml COMMAND ${PYTHON_EXECUTABLE} parse_api.py --api_yaml_path ./backward.yaml
--output_path ./parsed_apis/backward_api.parsed.yaml --backward --output_path ./parsed_apis/backward_api.parsed.yaml --backward
COMMAND COMMAND
${PYTHON_EXECUTABLE} parse_api.py --api_yaml_path ./new_backward.yaml ${PYTHON_EXECUTABLE} parse_api.py --api_yaml_path ./legacy_backward.yaml
--output_path ./parsed_apis/new_backward_api.parsed.yaml --backward --output_path ./parsed_apis/legacy_backward_api.parsed.yaml --backward
RESULTS_VARIABLE _results) RESULTS_VARIABLE _results)
foreach(_result in ${_results}) foreach(_result in ${_results})
if(${_result}) if(${_result})
...@@ -141,9 +141,9 @@ execute_process( ...@@ -141,9 +141,9 @@ execute_process(
WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen
COMMAND COMMAND
${PYTHON_EXECUTABLE} cross_validate.py --forward_yaml_paths ${PYTHON_EXECUTABLE} cross_validate.py --forward_yaml_paths
./parsed_apis/api.parsed.yaml ./parsed_apis/new_api.parsed.yaml ./parsed_apis/api.parsed.yaml ./parsed_apis/legacy_api.parsed.yaml
--backward_yaml_paths ./parsed_apis/backward_api.parsed.yaml --backward_yaml_paths ./parsed_apis/backward_api.parsed.yaml
./parsed_apis/new_backward_api.parsed.yaml ./parsed_apis/legacy_backward_api.parsed.yaml
RESULT_VARIABLE _result) RESULT_VARIABLE _result)
if(${_result}) if(${_result})
message(FATAL_ERROR "api validation failed, exiting.") message(FATAL_ERROR "api validation failed, exiting.")
...@@ -158,8 +158,8 @@ execute_process( ...@@ -158,8 +158,8 @@ execute_process(
WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}/python/paddle/utils/code_gen
COMMAND COMMAND
${PYTHON_EXECUTABLE} generate_op.py --api_yaml_path ${PYTHON_EXECUTABLE} generate_op.py --api_yaml_path
./parsed_apis/new_api.parsed.yaml --backward_api_yaml_path ./parsed_apis/api.parsed.yaml --backward_api_yaml_path
./parsed_apis/new_backward_api.parsed.yaml --output_op_path ./parsed_apis/backward_api.parsed.yaml --output_op_path
"${generated_op_path}.tmp" --output_arg_map_path "${generated_op_path}.tmp" --output_arg_map_path
"${generated_argument_mapping_path}.tmp" "${generated_argument_mapping_path}.tmp"
RESULT_VARIABLE _result) RESULT_VARIABLE _result)
...@@ -209,7 +209,7 @@ add_custom_command( ...@@ -209,7 +209,7 @@ add_custom_command(
COMMAND ${PYTHON_EXECUTABLE} -m pip install pyyaml COMMAND ${PYTHON_EXECUTABLE} -m pip install pyyaml
COMMAND COMMAND
${PYTHON_EXECUTABLE} ${api_gen_file} --api_yaml_path ${api_yaml_file} ${PYTHON_EXECUTABLE} ${api_gen_file} --api_yaml_path ${api_yaml_file}
${new_api_yaml_file} --api_header_path ${api_header_file_tmp} ${legacy_api_yaml_file} --api_header_path ${api_header_file_tmp}
--api_header_path ${api_header_file_tmp} --api_source_path --api_header_path ${api_header_file_tmp} --api_source_path
${api_source_file_tmp} ${api_source_file_tmp}
COMMAND ${CMAKE_COMMAND} -E copy_if_different ${api_header_file_tmp} COMMAND ${CMAKE_COMMAND} -E copy_if_different ${api_header_file_tmp}
...@@ -226,7 +226,7 @@ add_custom_command( ...@@ -226,7 +226,7 @@ add_custom_command(
${bw_api_source_file_tmp} ${bw_api_source_file_tmp}
COMMAND COMMAND
${PYTHON_EXECUTABLE} ${bw_api_gen_file} --backward_yaml_path ${PYTHON_EXECUTABLE} ${bw_api_gen_file} --backward_yaml_path
${bw_api_yaml_file} ${new_bw_api_yaml_file} --backward_header_path ${bw_api_yaml_file} ${legacy_bw_api_yaml_file} --backward_header_path
${bw_api_header_file_tmp} --backward_source_path ${bw_api_source_file_tmp} ${bw_api_header_file_tmp} --backward_source_path ${bw_api_source_file_tmp}
COMMAND ${CMAKE_COMMAND} -E copy_if_different ${bw_api_header_file_tmp} COMMAND ${CMAKE_COMMAND} -E copy_if_different ${bw_api_header_file_tmp}
${bw_api_header_file} ${bw_api_header_file}
...@@ -293,7 +293,7 @@ add_custom_command( ...@@ -293,7 +293,7 @@ add_custom_command(
OUTPUT ${dygraph_api_header_file} ${dygraph_api_source_file} OUTPUT ${dygraph_api_header_file} ${dygraph_api_source_file}
COMMAND COMMAND
${PYTHON_EXECUTABLE} ${im_api_gen_file} --api_yaml_path ${api_yaml_file} ${PYTHON_EXECUTABLE} ${im_api_gen_file} --api_yaml_path ${api_yaml_file}
${new_api_yaml_file} --sparse_api_yaml_path ${sparse_api_yaml_file} ${legacy_api_yaml_file} --sparse_api_yaml_path ${sparse_api_yaml_file}
--dygraph_api_header_path ${dygraph_api_header_file_tmp} --dygraph_api_header_path ${dygraph_api_header_file_tmp}
--dygraph_api_source_path ${dygraph_api_source_file_tmp} --dygraph_api_source_path ${dygraph_api_source_file_tmp}
COMMAND ${CMAKE_COMMAND} -E copy_if_different ${dygraph_api_header_file_tmp} COMMAND ${CMAKE_COMMAND} -E copy_if_different ${dygraph_api_header_file_tmp}
...@@ -309,7 +309,7 @@ add_custom_command( ...@@ -309,7 +309,7 @@ add_custom_command(
OUTPUT ${wrapped_infermeta_header_file} ${wrapped_infermeta_source_file} OUTPUT ${wrapped_infermeta_header_file} ${wrapped_infermeta_source_file}
COMMAND COMMAND
${PYTHON_EXECUTABLE} ${wrapped_infermeta_gen_file} --api_yaml_path ${PYTHON_EXECUTABLE} ${wrapped_infermeta_gen_file} --api_yaml_path
${api_yaml_file} ${new_api_yaml_file} --wrapped_infermeta_header_path ${api_yaml_file} ${legacy_api_yaml_file} --wrapped_infermeta_header_path
${wrapped_infermeta_header_file} --wrapped_infermeta_source_path ${wrapped_infermeta_header_file} --wrapped_infermeta_source_path
${wrapped_infermeta_source_file} ${wrapped_infermeta_source_file}
DEPENDS ${api_yaml_file} ${wrapped_infermeta_gen_file} ${api_gen_base} DEPENDS ${api_yaml_file} ${wrapped_infermeta_gen_file} ${api_gen_base}
......
python/paddle/utils/code_gen/api.yaml
浏览文件 @ fcd32950
- api : abs
args : (Tensor x)
output : Tensor
infer_meta :
func : RealAndImagInferMeta
kernel :
func : abs
backward : abs_grad
# accuracy
- api : accuracy
args : (Tensor x, Tensor indices, Tensor label)
output : Tensor(accuracy), Tensor(correct), Tensor(total)
infer_meta :
func : AccuracyInferMeta
kernel :
func : accuracy
dtype : x
# acos
- api : acos
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : acos
backward : acos_grad
# acosh
- api : acosh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : acosh
backward : acosh_grad
- api : adadelta
args : (Tensor param, Tensor grad, Tensor avg_squared_grad, Tensor avg_squared_update, float rho, float epsilon)
output : Tensor(param_out), Tensor(moment_out), Tensor(inf_norm_out)
infer_meta :
func : AdadeltaInferMeta
kernel :
func : adadelta
- api : adam
args : (Tensor param, Tensor grad, Tensor learning_rate, Tensor moment1, Tensor moment2, Tensor beta1_pow, Tensor beta2_pow, Tensor master_param, Tensor skip_update, Scalar beta1, Scalar beta2, Scalar epsilon, bool lazy_mode, int64_t min_row_size_to_use_multithread, bool multi_precision, bool use_global_beta_pow)
output : Tensor(param_out), Tensor(moment1_out), Tensor(moment2_out), Tensor(beta1_pow_out), Tensor(beta2_pow_out), Tensor(master_param_outs)
optional : master_param, skip_update
invoke : adam_impl(param, grad, learning_rate, moment1, moment2, beta1_pow, beta2_pow, master_param, skip_update, beta1, beta2, epsilon, lazy_mode, min_row_size_to_use_multithread, multi_precision, use_global_beta_pow)
- api : adamax
args : (Tensor param, Tensor grad, Tensor learning_rate, Tensor moment, Tensor inf_norm, Tensor beta1_pow, float beta1, float beta2, float epsilon)
output : Tensor(param_out), Tensor(avg_squared_grad_out), Tensor(avg_squared_update_out)
infer_meta :
func : AdamaxInferMeta
kernel :
func : adamax
- api : adamw
args : (Tensor param, Tensor grad, Tensor learning_rate, Tensor moment1, Tensor moment2, Tensor beta1_pow, Tensor beta2_pow, Tensor master_param, Tensor skip_update, Scalar beta1, Scalar beta2, Scalar epsilon, float lr_ratio, float coeff, bool with_decay, bool lazy_mode, int64_t min_row_size_to_use_multithread, bool multi_precision, bool use_global_beta_pow)
output : Tensor(param_out), Tensor(moment1_out), Tensor(moment2_out), Tensor(beta1_pow_out), Tensor(beta2_pow_out), Tensor(master_param_outs)
optional : master_param, skip_update
invoke : adamw_impl(param, grad, learning_rate, moment1, moment2, beta1_pow, beta2_pow, master_param, skip_update, beta1, beta2, epsilon, lr_ratio, coeff, with_decay, lazy_mode, min_row_size_to_use_multithread, multi_precision, use_global_beta_pow)
- api : add
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : add
backward : add_grad
- api : add_n
args : (Tensor[] x)
output : Tensor
infer_meta :
func : AddNInferMeta
kernel :
func : add_n
backward : add_n_grad
- api : addmm
args : (Tensor input, Tensor x, Tensor y, float alpha, float beta)
output : Tensor
infer_meta :
func : AddmmInferMeta
kernel :
func : addmm
backward : addmm_grad
- api : all
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : all
- api : allclose
args : (Tensor x, Tensor y, Scalar rtol, Scalar atol, bool equal_nan)
output : Tensor(out)
infer_meta :
func : AllValueCompareInferMeta
param: [x, y]
kernel :
func : allclose
- api : any
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : any
- api : arange
args : (Tensor start, Tensor end, Tensor step, DataType dtype, Place place={})
output : Tensor
infer_meta :
func : ArangeInferMeta
param : [start, end, step]
kernel :
func : arange
param : [start, end, step]
data_type : dtype
backend : place
data_transform :
support_trans_dtype : start, end, step
# arg_max
- api : argmax
args : (Tensor x, int64_t axis, bool keepdims, bool flatten, int dtype)
output : Tensor
infer_meta :
func : ArgMinMaxInferMeta
kernel :
func : arg_max
# arg_min
- api : argmin
args : (Tensor x, int64_t axis, bool keepdims, bool flatten, int dtype)
output : Tensor
infer_meta :
func : ArgMinMaxInferMeta
kernel :
func : arg_min
# argsort
- api : argsort
args : (Tensor x, int axis, bool descending)
output : Tensor(out), Tensor(indices)
infer_meta :
func : ArgsortInferMeta
kernel :
func : argsort
backward : argsort_grad
# asin
- api : asin
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : asin
backward : asin_grad
# asinh
- api : asinh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : asinh
backward : asinh_grad
# assign
- api : assign
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : assign
backward : assign_grad
- api : assign_out_
args : (Tensor x, Tensor output)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : assign
param : [x]
inplace : (output -> out)
backward : assign_out__grad
# atan
- api : atan
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : atan
backward : atan_grad
- api : atan2
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : Atan2InferMeta
kernel :
func : atan2
backward : atan2_grad
# atanh
- api : atanh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : atanh
backward : atanh_grad
# auc
- api : auc
args : (Tensor x, Tensor label, Tensor stat_pos, Tensor stat_neg, str curve, int num_thresholds, int slide_steps)
output : Tensor(auc), Tensor(stat_pos_out), Tensor(stat_neg_out)
infer_meta :
func : AucInferMeta
kernel :
func : auc
# batch_norm
- api : batch_norm
args : (Tensor x, Tensor scale, Tensor bias, Tensor mean, Tensor variance, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu)
output : Tensor(out), Tensor(mean_out), Tensor(variance_out), Tensor(saved_mean), Tensor(saved_variance), Tensor(reserve_space)
invoke : batch_norm_impl(x, scale, bias, mean, variance, momentum, epsilon, data_layout, is_test, use_global_stats, trainable_statistics, fuse_with_relu)
backward : batch_norm_grad
- api : bce_loss
args : (Tensor input, Tensor label)
output : Tensor
infer_meta :
func : BCELossInferMeta
kernel :
func : bce_loss
backward : bce_loss_grad
# bernoulli
- api : bernoulli
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : bernoulli
# bitwise_and
- api : bitwise_and
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : bitwise_and
# bitwise_not
- api : bitwise_not
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : bitwise_not
# bitwise_or
- api : bitwise_or
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : bitwise_or
# bitwise_xor
- api : bitwise_xor
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : bitwise_xor
# brelu
- api : brelu
args : (Tensor x, float t_min, float t_max)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : brelu
backward : brelu_grad
- api : cast
args : (Tensor x, DataType out_dtype)
output : Tensor
infer_meta :
func : CastInferMeta
kernel :
func : cast
param : [x, out_dtype]
data_type : x
backward : cast_grad
- api : ceil
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : ceil
backward : ceil_grad
- api : celu
args : (Tensor x, float alpha)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : celu
backward : celu_grad
# cholesky
- api : cholesky
args : (Tensor x, bool upper)
output : Tensor
infer_meta :
func : CholeskyInferMeta
kernel :
func : cholesky
backward : cholesky_grad
# cholesky_solve
- api : cholesky_solve
args : (Tensor x, Tensor y, bool upper)
output : Tensor
infer_meta :
func : CholeskySolveInferMeta
kernel :
func : cholesky_solve
backward : cholesky_solve_grad
- api : clip
args : (Tensor x, Scalar(float) min, Scalar(float) max)
output : Tensor(out)
inplace : (x -> out)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : clip
backward : clip_grad
- api : concat
args : (Tensor[] x, Scalar(int64_t) axis)
output : Tensor
infer_meta :
func : ConcatInferMeta
param : [x, axis]
kernel :
func : concat
backward : concat_grad
- api : conj
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : conj
backward : conj_grad
- api : conv2d
args : (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor
invoke : conv2d_impl(input, filter, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search)
backward : conv2d_grad
- api : conv2d_transpose
args : (Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(out)
infer_meta :
func : ConvTransposeInferMeta
kernel :
func : conv2d_transpose
use_gpudnn : true
backward : conv2d_transpose_grad
- api : conv3d
args : (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor
invoke : conv3d_impl(input, filter, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search)
backward : conv3d_grad
- api : conv3d_transpose
args : (Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(out)
infer_meta :
func : ConvTransposeInferMeta
kernel :
func : conv3d_transpose
use_gpudnn : true
backward : conv3d_transpose_grad
- api : copy_to
args : (Tensor x, Place place, bool blocking)
output : Tensor
invoke : copy_to_impl(x, place, blocking)
# cos
- api : cos
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : cos
backward : cos_grad
# cosh
- api : cosh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : cosh
backward : cosh_grad
- api : cross
args : (Tensor x, Tensor y, int axis = 9)
output : Tensor
infer_meta :
func : CrossInferMeta
kernel :
func : cross
backward : cross_grad
# Part of python API paddle.nn.functional.cross_entropy
- api : cross_entropy_with_softmax
args : (Tensor input, Tensor label, bool soft_label, bool use_softmax, bool numeric_stable_mode, int ignore_index, int axis)
output : Tensor(softmax), Tensor(loss)
infer_meta :
func : CrossEntropyWithSoftmaxInferMeta
kernel :
func : cross_entropy_with_softmax
data_type : input
backward : cross_entropy_with_softmax_grad
- api : cumprod
args : (Tensor x, int dim)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : cumprod
backward : cumprod_grad
- api : cumsum
args : (Tensor x, int axis, bool flatten, bool exclusive, bool reverse)
output : Tensor(out)
infer_meta :
func : CumInferMeta
kernel :
func : cumsum
backward : cumsum_grad
- api : deformable_conv
args : (Tensor x, Tensor offset, Tensor filter, Tensor mask, int[] strides, int[] paddings, int[] dilations, int deformable_groups, int groups, int im2col_step)
output : Tensor(out)
infer_meta :
func : DeformableConvInferMeta
kernel :
func : deformable_conv
data_type : x
optional : mask
backward : deformable_conv_grad
- api : depthwise_conv2d
args : (Tensor x, Tensor filter, int[] strides, int[] paddings, str padding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn)
output : Tensor(out)
infer_meta :
func : ConvInferMeta
param : [x, filter, strides, paddings, padding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search]
kernel :
func : depthwise_conv2d
param : [x, filter, strides, paddings, padding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search, fuse_relu]
use_gpudnn : use_gpudnn
backward : depthwise_conv2d_grad
- api : depthwise_conv2d_transpose
args : (Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(out)
infer_meta :
func : ConvTransposeInferMeta
kernel :
func : depthwise_conv2d_transpose
backward : depthwise_conv2d_transpose_grad
- api : det
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : determinant
backward : det_grad
- api : diag
args : (Tensor x, int offset, float padding_value)
output : Tensor
infer_meta :
func : DiagInferMeta
kernel :
func : diag
- api : diagonal
args : (Tensor x, int offset, int axis1, int axis2)
output : Tensor
infer_meta :
func : DiagonalInferMeta
kernel :
func : diagonal
backward : diagonal_grad
- api : digamma
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : digamma
backward : digamma_grad
- api : dist
args : (Tensor x, Tensor y, float p)
output : Tensor
infer_meta :
func : DistInferMeta
kernel :
func : dist
backward : dist_grad
- api : divide
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : divide
backward : divide_grad
- api : dot
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : DotInferMeta
kernel :
func : dot
- api : dropout
args : (Tensor x, Tensor seed_tensor, float p, bool is_test, str mode, int seed, bool fix_seed)
output : Tensor(out), Tensor(mask)
infer_meta :
func : DropoutInferMeta
kernel :
func : dropout
data_type : x
optional : seed_tensor
backward : dropout_grad
# eigh
- api : eigh
args : (Tensor x, str uplo)
output : Tensor(out_w), Tensor(out_v)
infer_meta :
func : EighInferMeta
kernel :
func : eigh
backward : eigh_grad
- api : einsum
args : (Tensor[] x, str equation)
output : Tensor, Tensor[]{x.size()}, Tensor[]{x.size()}
infer_meta :
func : EinsumInferMeta
param : [x, equation]
kernel :
func : einsum
backward : einsum_grad
- api : elementwise_pow
args : (Tensor x, Tensor y)
output : Tensor(out)
infer_meta :
func : ElementwiseInferMeta
kernel :
func : elementwise_pow
backward : elementwise_pow_grad
# elu
- api : elu
args : (Tensor x, float alpha)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : elu
backward : elu_grad
- api : embedding
args : (Tensor x, Tensor weight, int64_t padding_idx=-1, bool sparse=false)
output : Tensor
invoke : embedding_impl(x, weight, padding_idx, sparse)
backward : embedding_grad
- api : empty
args : (IntArray shape, DataType dtype=DataType::FLOAT32, Place place=CPUPlace())
output: Tensor
infer_meta :
func : CreateInferMeta
param : [shape, dtype]
kernel :
func : empty
param : [shape, dtype]
data_type : dtype
backend : place
- api : empty_like
args : (Tensor x, DataType dtype = DataType::UNDEFINED, Place place = {})
output: Tensor
infer_meta :
func : CreateLikeInferMeta
param : [x, dtype]
kernel :
func : empty_like
param : [x, dtype]
data_type : dtype > x
backend : place > x
- api : equal
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : equal
- api : equal_all
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : CompareAllInferMeta
kernel :
func : equal_all
# erf
- api : erf
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : erf
backward : erf_grad
# erfinv
- api : erfinv
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : erfinv
backward : erfinv_grad
# exp
- api : exp
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : exp
backward : exp_grad
# expand
- api : expand
args : (Tensor x, IntArray shape)
output : Tensor
infer_meta :
func : ExpandInferMeta
kernel :
func : expand
backward : expand_grad
# expand_as
- api : expand_as
args : (Tensor x, Tensor y, int[] target_shape)
output : Tensor
infer_meta :
func : ExpandAsInferMeta
kernel :
func : expand_as
optional : y
backward : expand_as_grad
- api : expm1
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : expm1
backward : expm1_grad
- api : eye
args : (int64_t num_rows, int64_t num_columns, DataType dtype=DataType::FLOAT32, Place place={})
output : Tensor(out)
infer_meta :
func : EyeInferMeta
param : [num_rows, num_columns, dtype]
kernel :
func : eye
param : [num_rows, num_columns, dtype]
data_type : dtype
backend : place
- api : flatten
args : (Tensor x, int start_axis, int stop_axis)
output : Tensor(out), Tensor(xshape)
infer_meta :
func : FlattenWithXShapeInferMeta
kernel :
func : flatten_with_xshape
backend : x
inplace : (x -> out)
view : (x -> out)
intermediate : xshape
backward : flatten_grad
# flip
- api : flip
args : (Tensor x, int[] axis)
output : Tensor
infer_meta :
func : FlipInferMeta
kernel :
func : flip
backward : flip_grad
- api : floor
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : floor
backward : floor_grad
- api : floor_divide
args : (Tensor x, Tensor y)
output : Tensor(out)
infer_meta :
func : ElementwiseInferMeta
kernel :
func : floor_divide
- api : fmax
args : (Tensor x, Tensor y, int axis)
output : Tensor(out)
infer_meta :
param: [x, y]
func : ElementwiseInferMeta
kernel :
func : fmax
backward : fmax_grad
- api : fmin
args : (Tensor x, Tensor y, int axis)
output : Tensor(out)
infer_meta :
param: [x, y]
func : ElementwiseInferMeta
kernel :
func : fmin
backward : fmin_grad
- api : frobenius_norm
args : (Tensor x, int64_t[] axis, bool keep_dim, bool reduce_all)
output : Tensor(out)
infer_meta :
func : ReduceInferMetaBase
kernel :
func : frobenius_norm
backward : frobenius_norm_grad
- api : full
args : (IntArray shape, Scalar value, DataType dtype=DataType::FLOAT32, Place place=CPUPlace())
output: Tensor
infer_meta :
func : CreateInferMeta
param : [shape, dtype]
kernel :
func : full
param : [shape, value, dtype]
data_type : dtype
backend : place
- api : full_batch_size_like
args : (Tensor input, int[] shape, DataType dtype, Scalar value, int input_dim_idx, int output_dim_idx, Place place=CPUPlace())
output: Tensor
infer_meta :
func : FullBatchSizeLikeInferMeta
param : [input, shape, value, dtype, input_dim_idx, output_dim_idx]
kernel :
func : full_batch_size_like
param : [input, shape, value, dtype, input_dim_idx, output_dim_idx]
data_type : dtype
backend : place
- api : full_like
args : (Tensor x, Scalar value, DataType dtype = DataType::UNDEFINED, Place place = {})
output: Tensor
infer_meta :
func : CreateLikeInferMeta
param : [x, dtype]
kernel :
func : full_like
param : [x, value, dtype]
data_type : dtype > x
backend : place > x
data_transform :
skip_transform : x
- api : gather
args : (Tensor x, Tensor index, Scalar(int) axis=0)
output : Tensor(out)
infer_meta :
func : GatherInferMeta
kernel :
func : gather
data_type: x
backward : gather_grad
- api : gather_nd
args : (Tensor x, Tensor index)
output : Tensor
infer_meta :
func : GatherNdInferMeta
kernel :
func : gather_nd
data_type : x
backward : gather_nd_grad
- api : gather_tree
args : (Tensor ids, Tensor parents)
output : Tensor
infer_meta :
func : GatherTreeMeta
kernel :
func : gather_tree
- api : gaussian_random
args : (IntArray shape, float mean, float std, int seed, DataType dtype, Place place={})
output: Tensor
infer_meta :
func : GaussianRandomInferMeta
param : [shape, mean, std, seed, dtype]
kernel :
func : gaussian_random
param : [shape, mean, std, seed, dtype]
data_type : dtype
backend : place
- api : gelu
args : (Tensor x, bool approximate)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : gelu
backward : gelu_grad
- api : graph_send_recv
args : (Tensor x, Tensor src_index, Tensor dst_index, str pool_type = "SUM", int64_t out_size = 0)
output : Tensor(out), Tensor(dst_count)
infer_meta :
func : GraphSendRecvInferMeta
kernel :
func : graph_send_recv
data_type : x
intermediate : dst_count
backward : graph_send_recv_grad
- api : greater_equal
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : greater_equal
- api : greater_than
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : greater_than
- api : group_norm
args : (Tensor x, Tensor scale, Tensor bias, float epsilon, int groups, str data_layout)
output : Tensor(y), Tensor(mean), Tensor(variance)
infer_meta :
func : GroupNormInferMeta
kernel :
func : group_norm
optional : scale, bias
intermediate : mean, variance
backward : group_norm_grad
- api : gumbel_softmax
args : (Tensor x, float temperature, bool hard, int axis)
output : Tensor
infer_meta :
func : GumbelSoftmaxInferMeta
kernel :
func : gumbel_softmax
backward : gumbel_softmax_grad
# hard_shrink
- api : hard_shrink
args : (Tensor x, float threshold)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_shrink
backward : hard_shrink_grad
# hard_sigmoid
- api : hard_sigmoid
args : (Tensor x, float slope, float offset)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_sigmoid
backward : hard_sigmoid_grad
- api : hard_swish
args : (Tensor x, float threshold = 6.0, float scale = 6.0, float offset = 3.0)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_swish
backward : hard_swish_grad
# histogram
- api : histogram
args : (Tensor x, int64_t bins, int min, int max)
output : Tensor
infer_meta :
func : HistogramInferMeta
kernel :
func : histogram
- api : huber_loss
args : (Tensor input, Tensor label, float delta)
output : Tensor(out), Tensor(residual)
infer_meta :
func : HuberLossInferMeta
kernel :
func : huber_loss
backward : huber_loss_grad
- api : imag
args : (Tensor x)
output : Tensor
infer_meta :
func : RealAndImagInferMeta
kernel :
func : imag
backward : imag_grad
# increment
- api : increment
args : (Tensor x, float value)
output : Tensor
infer_meta :
func : IncrementInferMeta
kernel :
func : increment
- api : index_sample
args : (Tensor x, Tensor index)
output : Tensor
infer_meta :
func : IndexSampleInferMeta
kernel :
func : index_sample
data_type : x
backward : index_sample_grad
- api : index_select
args : (Tensor x, Tensor index, int dim)
output : Tensor(out)
infer_meta :
func : IndexSelectInferMeta
kernel :
func : index_select
data_type : x
backward : index_select_grad
- api : instance_norm
args : (Tensor x, Tensor scale, Tensor bias, float epsilon)
output : Tensor(y), Tensor(saved_mean), Tensor(saved_variance)
infer_meta :
func : InstanceNormInferMeta
kernel :
func : instance_norm
data_type : x
optional : scale, bias
intermediate : saved_mean, saved_variance
backward : instance_norm_grad
# is_empty
- api : is_empty
args : (Tensor x)
output : Tensor
infer_meta :
func : IsEmptyInferMeta
kernel :
func : is_empty
- api : isclose
args : (Tensor x, Tensor y, Scalar rtol, Scalar atol, bool equal_nan)
output : Tensor(out)
infer_meta :
func : ValueCompareInferMeta
param: [x, y]
kernel :
func : isclose
# isfinite
- api : isfinite
args : (Tensor x)
output : Tensor
infer_meta :
func : IsfiniteInferMeta
kernel :
func : isfinite, infinite_sr
# isinf
- api : isinf
args : (Tensor x)
output : Tensor
infer_meta :
func : IsfiniteInferMeta
kernel :
func : isinf, isinf_sr
# isnan
- api : isnan
args : (Tensor x)
output : Tensor
infer_meta :
func : IsfiniteInferMeta
kernel :
func : isnan, isnan_sr
- api : kldiv_loss
args : (Tensor x, Tensor label, str reduction)
output : Tensor(out)
infer_meta :
func : KLDivInferMeta
kernel :
func : kldiv_loss
data_type : x
backward : kldiv_loss_grad
- api : kron
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : KronInferMeta
kernel :
func : kron
backward : kron_grad
- api : kthvalue
args : (Tensor x, int k, int axis, bool keepdim)
output : Tensor(out), Tensor(indices)
infer_meta :
func : KthvalueInferMeta
kernel :
func : kthvalue
backward : kthvalue_grad
# label_smooth
- api : label_smooth
args : (Tensor label, Tensor prior_dist, float epsilon)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [label]
kernel :
func : label_smooth
data_type : label
optional : prior_dist
backward : label_smooth_grad
- api : layer_norm
args : (Tensor x, Tensor scale, Tensor bias, float epsilon, int begin_norm_axis, bool is_test)
output : Tensor(out), Tensor(mean), Tensor(variance)
infer_meta :
func : LayerNormInferMeta
kernel :
func : layer_norm
data_type : x
backward : layer_norm_grad
optional : scale, bias
# leaky_relu
- api : leaky_relu
args : (Tensor x, float alpha)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : leaky_relu
backward : leaky_relu_grad
- api : lerp
args : (Tensor x, Tensor y, Tensor weight)
output : Tensor
infer_meta :
func : LerpInferMeta
kernel :
func : lerp
backward : lerp_grad
- api : less_equal
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : less_equal
- api : less_than
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : less_than
- api : lgamma
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : lgamma
backward : lgamma_grad
- api : linspace
args : (Tensor start, Tensor stop, Tensor number, DataType dtype)
output : Tensor
infer_meta :
func : LinspaceInferMeta
kernel :
func : linspace
data_type : dtype
- api : log
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : log
backward: log_grad
- api : log10
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : log10
backward: log10_grad
- api : log1p
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : log1p
backward: log1p_grad
- api : log2
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : log2
backward: log2_grad
# log_loss
- api : log_loss
args : (Tensor input, Tensor label, float epsilon)
output : Tensor
infer_meta :
func : LogLossInferMeta
kernel :
func : log_loss
backward : log_loss_grad
- api : log_softmax
args : (Tensor x, int axis)
output : Tensor(out)
infer_meta :
func : UnchangedInferMetaCheckAxis
kernel :
func : log_softmax
backward : log_softmax_grad
- api : logcumsumexp
args : (Tensor x, int axis, bool flatten, bool exclusive, bool reverse)
output : Tensor(out)
infer_meta :
func : CumInferMeta
kernel :
func : logcumsumexp
backward : logcumsumexp_grad
# logical_and
- api : logical_and
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : logical_and
# logical_not
- api : logical_not
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : logical_not
# logical_or
- api : logical_or
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : logical_or
# logical_xor
- api : logical_xor
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : logical_xor
# logit
- api : logit
args : (Tensor x, float eps = 1e-6f)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : logit
backward : logit_grad
# logsigmoid
- api : logsigmoid
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : logsigmoid
backward : logsigmoid_grad
- api : logsumexp
args : (Tensor x, int64_t[] axis, bool keepdim, bool reduce_all)
output : Tensor(out)
infer_meta :
func : LogsumexpInferMeta
kernel :
func : logsumexp
backward : logsumexp_grad
# masked_select
- api : masked_select
args : (Tensor x, Tensor mask)
output : Tensor
infer_meta :
func : MaskedSelectInferMeta
kernel :
func : masked_select
data_type : x
backward : masked_select_grad
- api : matmul
args : (Tensor x, Tensor y, bool transpose_x = false, bool transpose_y = false)
output : Tensor
infer_meta :
func : MatmulInferMeta
kernel :
func : matmul
backward : matmul_grad
# matrix_power
- api : matrix_power
args : (Tensor x, int n)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : matrix_power
backward : matrix_power_grad
- api : matrix_rank
args : (Tensor x, float tol, bool use_default_tol=true, bool hermitian=false)
output : Tensor(out)
infer_meta :
func : MatrixRankInferMeta
param : [x, use_default_tol, hermitian]
kernel :
func : matrix_rank
- api : matrix_rank_tol
args : (Tensor x, Tensor atol_tensor, bool use_default_tol=true, bool hermitian=false)
output : Tensor(out)
infer_meta :
func : MatrixRankTolInferMeta
kernel :
func : matrix_rank_tol
- api : max
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : max
backward : max_grad
- api : max_pool2d_with_index
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
output : Tensor(out), Tensor(mask)
infer_meta :
func : MaxPoolWithIndexInferMeta
kernel :
func : max_pool2d_with_index
backward : max_pool2d_with_index_grad
- api : max_pool3d_with_index
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
output : Tensor(out), Tensor(mask)
infer_meta :
func : MaxPoolWithIndexInferMeta
kernel :
func : max_pool3d_with_index
backward : max_pool3d_with_index_grad
- api : maximum
args : (Tensor x, Tensor y)
output : Tensor(out)
infer_meta :
func : ElementwiseInferMeta
kernel :
func : maximum
backward : maximum_grad
- api : maxout
args : (Tensor x, int groups, int axis)
output : Tensor(out)
infer_meta :
func : MaxOutInferMeta
kernel :
func : maxout
backward : maxout_grad
- api : mean
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : mean
backward : mean_grad
- api : mean_all
args : (Tensor x)
output : Tensor
infer_meta :
func : MeanAllInferMeta
kernel :
func : mean_all
backward : mean_all_grad
- api : meshgrid
args : (Tensor[] inputs)
output : Tensor[]{inputs.size()}
infer_meta :
func : MeshgridInferMeta
kernel :
func : meshgrid
backward : meshgrid_grad
- api : min
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : min
backward : min_grad
- api : minimum
args : (Tensor x, Tensor y)
output : Tensor(out)
infer_meta :
func : ElementwiseInferMeta
kernel :
func : minimum
backward : minimum_grad
- api : mish
args : (Tensor x, float lambda)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : mish
backward : mish_grad
- api : mode
args : (Tensor x, int axis, bool keepdim)
output : Tensor(out), Tensor(indices)
infer_meta :
func : ModeInferMeta
kernel :
func : mode
backward : mode_grad
- api : modulo
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : modulo
backward : modulo_grad
- api : momentum
args : (Tensor param, Tensor grad, Tensor velocity, Tensor learning_rate, Tensor master_param, float mu, bool use_nesterov = false, str regularization_method = "", float regularization_coeff = 0.0, bool multi_precision = false, float rescale_grad = 1.0f)
output : Tensor(param_out), Tensor(velocity_out), Tensor(master_param_out)
invoke : momentum_impl(param, grad, velocity, learning_rate, master_param, mu, use_nesterov, regularization_method, regularization_coeff, multi_precision, rescale_grad)
optional : master_param
- api : multi_dot
args : (Tensor[] x)
output : Tensor
infer_meta :
func : MultiDotInferMeta
kernel :
func : multi_dot
backward : multi_dot_grad
# multinomial
- api : multinomial
args : (Tensor x, int num_samples, bool replacement)
output : Tensor
infer_meta :
func : MultinomialInferMeta
kernel :
func : multinomial
- api : multiplex
args : (Tensor[] ins, Tensor ids)
output : Tensor
infer_meta :
func : MultiplexInferMeta
kernel :
func : multiplex
data_type : ins
backward : multiplex_grad
- api : multiply
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : multiply
backward : multiply_grad
- api : mv
args : (Tensor x, Tensor vec)
output : Tensor
infer_meta :
func : MvInferMeta
kernel :
func : mv
backward : mv_grad
- api : nll_loss
args : (Tensor input, Tensor label, Tensor weight, int64_t ignore_index, str reduction)
output : Tensor(out), Tensor(total_weight)
infer_meta :
func : NllLossRawInferMeta
kernel :
func : nll_loss
data_type : input
optional : weight
backward : nll_loss_grad
- api : norm
args : (Tensor x, int axis, float epsilon, bool is_test)
output : Tensor(out), Tensor(norm)
infer_meta :
func : NormInferMeta
kernel :
func : norm
intermediate : norm
backward : norm_grad
- api : not_equal
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : not_equal
- api : one_hot
args : (Tensor x, Scalar(int) num_classes)
output : Tensor
infer_meta :
func : OneHotInferMeta
kernel :
func : one_hot
- api : ones_like
args : (Tensor x, DataType dtype=DataType::UNDEFINED, Place place={})
output : Tensor
invoke : full_like(x, 1, dtype, place)
- api : p_norm
args : (Tensor x, float porder, int axis, float epsilon, bool keepdim, bool asvector=false)
output : Tensor(out)
infer_meta :
func : PNormInferMeta
kernel :
func : p_norm
backward : p_norm_grad
# pad
- api : pad
args : (Tensor x, int[] paddings, float pad_value)
output : Tensor
infer_meta :
func : PadInferMeta
kernel :
func : pad
backward : pad_grad
- api : pad3d
args : (Tensor x, IntArray paddings, str mode, float pad_value, str data_format)
output : Tensor(out)
infer_meta :
func : Pad3dInferMeta
kernel :
func : pad3d
backward : pad3d_grad
# pixel_shuffle
- api : pixel_shuffle
args : (Tensor x, int upscale_factor, str data_format)
output : Tensor
infer_meta :
func : PixelShuffleInferMeta
kernel :
func : pixel_shuffle
backward : pixel_shuffle_grad
# poisson
- api : poisson
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : poisson
backward : poisson_grad
- api : pool2d
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(out)
infer_meta :
func : PoolInferMeta
kernel :
func : pool2d
use_gpudnn : true
backward : pool2d_grad
# Used in adaptive_avg_pool2d API
- api : pool2d_gpudnn_unused
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(out)
infer_meta :
func : PoolInferMeta
kernel :
func : pool2d
use_gpudnn : false
backward : pool2d_grad_gpudnn_unused
- api : pool3d
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(out)
infer_meta :
func : PoolInferMeta
kernel :
func : pool3d
use_gpudnn : true
backward : pool3d_grad
- api : pow
args : (Tensor x, Scalar s)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : pow
backward : pow_grad
- api : prelu
args : (Tensor x, Tensor alpha, str data_format, str mode)
output : Tensor(out)
infer_meta :
func : PReluInferMeta
kernel :
func : prelu
backward : prelu_grad
- api : psroi_pool
args : (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, int output_channels, float spatial_scale)
output : Tensor
infer_meta :
func : PsroiPoolInferMeta
kernel :
func : psroi_pool
data_type : x
optional : boxes_num
backward : psroi_pool_grad
# put_along_axis
- api : put_along_axis
args : (Tensor x, Tensor index, Tensor value, int axis, str reduce)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [index]
kernel :
func : put_along_axis
data_type : x
backward : put_along_axis_grad
- api : qr
args : (Tensor x, str mode)
output : Tensor(q), Tensor(r)
infer_meta :
func : QrInferMeta
kernel :
func : qr
# backward : qr_grad
- api : randint
args : (int low, int high, IntArray shape, DataType dtype=DataType::INT64, Place place={})
output : Tensor(out)
infer_meta :
func : RandintInferMeta
param : [low, high, shape, dtype]
kernel :
func : randint
param : [low, high, shape, dtype]
data_type : dtype
backend : place
- api : randperm
args : (int n, DataType dtype, Place place={})
output : Tensor
infer_meta :
func : RandpermInferMeta
param : [n, dtype]
kernel :
func : randperm
param : [n, dtype]
data_type : dtype
backend : place
- api : real
args : (Tensor x)
output : Tensor
infer_meta :
func : RealAndImagInferMeta
kernel :
func : real
backward : real_grad
- api : reciprocal
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : reciprocal
backward : reciprocal_grad
# reduce_prod
- api : reduce_prod
args : (Tensor x, int64_t[] dims, bool keep_dim, bool reduce_all)
output : Tensor
infer_meta :
func : ReduceInferMetaBase
kernel :
func : prod_raw
backward : reduce_prod_grad
- api : relu
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : relu
inplace : (x -> out)
backward : relu_grad
- api : reshape
args : (Tensor x, IntArray shape)
output : Tensor(out), Tensor(xshape)
infer_meta :
func : ReshapeWithXShapeInferMeta
kernel :
func : reshape_with_xshape
inplace : (x -> out)
view: (x -> out)
intermediate : xshape
backward: reshape_grad
- api : roi_align
args : (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale, int sampling_ratio, bool aligned)
output : Tensor
infer_meta :
func : RoiAlignInferMeta
kernel :
func : roi_align
data_type : x
optional : boxes_num
backward : roi_align_grad
- api : roi_pool
args : (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale)
output : Tensor(out), Tensor(arg_max)
infer_meta :
func : RoiPoolInferMeta
kernel :
func : roi_pool
data_type : x
optional : boxes_num
intermediate : arg_max
backward : roi_pool_grad
- api : roll
args : (Tensor x, IntArray shifts, int64_t[] axis)
output : Tensor(out)
infer_meta :
func : RollInferMeta
kernel :
func : roll
backward : roll_grad
- api : round
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : round
backward : round_grad
- api : rsqrt
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : rsqrt
inplace : (x -> out)
backward : rsqrt_grad
- api : scale
args : (Tensor x, Scalar scale, float bias, bool bias_after_scale)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : scale, scale_sr
inplace : (x -> out)
backward : scale_grad
- api : scatter
args : (Tensor x, Tensor index, Tensor updates, bool overwrite)
output : Tensor
infer_meta :
func : ScatterInferMeta
dtype : x
kernel :
func : scatter
backward : scatter_grad
- api : scatter_nd_add
args : (Tensor x, Tensor index, Tensor updates)
output : Tensor
infer_meta :
func : ScatterNdAddInferMeta
dtype : x
kernel :
func : scatter_nd_add
backward : scatter_nd_add_grad
- api : searchsorted
args : (Tensor sorted_sequence, Tensor value, bool out_int32, bool right)
output : Tensor(out)
infer_meta :
func : SearchsortedInferMeta
kernel :
func : searchsorted
data_type : sorted_sequence
# segment_pool
- api : segment_pool
args : (Tensor x, Tensor segment_ids, str pooltype)
output : Tensor(out), Tensor(summed_ids)
infer_meta :
func : SegmentPoolInferMeta
kernel :
func : segment_pool
data_type : x
backward : segment_pool_grad
# selu
- api : selu
args : (Tensor x, float scale, float alpha)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : selu
backward : selu_grad
- api : sgd
args : (Tensor param, Tensor learning_rate, Tensor grad, Tensor master_param, bool multi_precision)
output : Tensor(param_out), Tensor(master_param_out)
invoke : sgd_impl(param, learning_rate, grad, master_param, multi_precision)
optional : master_param
- api : shape
args : (Tensor input)
output : Tensor
infer_meta :
func : ShapeInferMeta
kernel :
func : shape, shape_sr
data_transform:
skip_transform : input
# shard_index
- api : shard_index
args : (Tensor in, int index_num, int nshards, int shard_id, int ignore_value)
output : Tensor
infer_meta :
func : ShardIndexInferMeta
kernel :
func : shard_index
# sigmoid
- api : sigmoid
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sigmoid
backward : sigmoid_grad
# sigmoid_cross_entropy_with_logits
- api : sigmoid_cross_entropy_with_logits
args : (Tensor x, Tensor label, bool normalize, int ignore_index)
output : Tensor
infer_meta :
func : SigmoidCrossEntropyWithLogitsInferMeta
kernel :
func : sigmoid_cross_entropy_with_logits
backward : sigmoid_cross_entropy_with_logits_grad
- api : sign
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sign
# silu
- api : silu
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : silu
backward : silu_grad
# sin
- api : sin
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sin
backward : sin_grad
# sinh
- api : sinh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sinh
backward : sinh_grad
# size
- api : size
args : (Tensor x)
output : Tensor
infer_meta :
func : SizeInferMeta
kernel :
func : size
data_transform:
skip_transform : x
- api : slice
args : (Tensor input, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis)
output : Tensor
infer_meta :
func : SliceRawInferMeta
kernel :
func : slice
backward : slice_grad
# soft_shrink
- api : soft_shrink
args : (Tensor x, float lambda)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : soft_shrink
backward : soft_shrink_grad
- api : softmax
args : (Tensor x, int axis)
output : Tensor
infer_meta :
func : SoftmaxInferMeta
kernel :
func : softmax
use_gpudnn : true
backward : softmax_grad
- api : split
args : (Tensor x, IntArray num_or_sections, Scalar(int) axis)
output : Tensor[]
invoke : split_impl(x, num_or_sections, axis)
backward : split_grad
- api : sqrt
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sqrt
backward : sqrt_grad
- api : square
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : square
backward : square_grad
- api : squeeze
args : (Tensor x, int[] axes)
output : Tensor(out), Tensor(xshape)
infer_meta :
func : SqueezeInferMeta
kernel :
func : squeeze
view: (x -> out)
intermediate : xshape
backward : squeeze_grad
- api : stack
args : (Tensor[] x, int axis)
output : Tensor
infer_meta :
func : StackInferMeta
kernel :
func : stack
backward : stack_grad
- api : strided_slice
args : (Tensor x, int[] axes, IntArray starts, IntArray ends, IntArray strides)
output : Tensor
infer_meta :
func : StridedSliceInferMeta
kernel :
func : strided_slice
backward : strided_slice_grad
- api : subtract
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : subtract
backward : subtract_grad
- api : sum
args : (Tensor x, int64_t[] dims={}, DataType out_dtype=DataType::UNDEFINED, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : SumInferMeta
kernel :
func : sum
data_type : x
backward : sum_grad
# The python API paddle.nn.functional.swish has no `bete` argument, it may be removed later
- api : swish
args : (Tensor x, float beta=1.0)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : swish
backward : swish_grad
# take_along_axis
- api : take_along_axis
args : (Tensor x, Tensor index, int axis)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [index]
kernel :
func : take_along_axis
data_type : x
backward : take_along_axis_grad
# tan
- api : tan
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : tan
backward : tan_grad
# tanh
- api : tanh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : tanh
backward : tanh_grad
# tanh_shrink
- api : tanh_shrink
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : tanh_shrink
backward : tanh_shrink_grad
# thresholded_relu
- api : thresholded_relu
args : (Tensor x, float threshold)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : thresholded_relu
backward : thresholded_relu_grad
# tile
- api : tile
args : (Tensor x, IntArray repeat_times)
output : Tensor
infer_meta :
func : TileInferMeta
kernel :
func : tile
backward : tile_grad
- api : top_k
args : (Tensor x, Scalar k, int axis = -1, bool largest = true, bool sorted = true)
output : Tensor(out), Tensor(indices)
infer_meta :
func : TopKInferMeta
kernel :
func : top_k
backward : top_k_grad
- api : trace
args : (Tensor x, int offset, int axis1, int axis2)
output : Tensor
infer_meta :
func : TraceInferMeta
kernel :
func : trace
backward : trace_grad
- api : transpose
args : (Tensor x, int[] axis)
output : Tensor
infer_meta :
func : TransposeInferMeta
kernel :
func : transpose
backward : transpose_grad
- api : triangular_solve
args : (Tensor x, Tensor y, bool upper, bool transpose, bool unitriangular)
output : Tensor
infer_meta :
func : TriangularSolveInferMeta
kernel :
func : triangular_solve
backward : triangular_solve_grad
- api : tril_indices
args : (int rows, int cols, int offset, DataType dtype, Place place={})
output : Tensor(out)
infer_meta :
func : TrilIndicesInferMeta
param : [rows, cols, offset, dtype]
kernel :
func : tril_indices
param : [rows, cols, offset, dtype]
data_type : dtype
backend : place
- api : tril_triu
args : (Tensor x, int diagonal, bool lower)
output : Tensor(out)
infer_meta :
func : TrilTriuInferMeta
kernel :
func : tril_triu
backward : tril_triu_grad
- api : trunc
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : trunc
backward : trunc_grad
# python API: paddle.nn.initializer.TruncatedNormal
- api : truncated_gaussian_random
args : (int[] shape, float mean, float std, int seed, DataType dtype=DataType::FLOAT32, Place place={})
output : Tensor
infer_meta :
func : TruncatedGaussianRandomInferMeta
param : [shape, mean, std, seed, dtype]
kernel :
func : truncated_gaussian_random
param : [shape, mean, std, seed, dtype]
backend : place
data_type : dtype
- api : unbind
args : (Tensor input, int axis)
output : Tensor[] {axis<0 ? input.dims()[input.dims().size()+axis]:input.dims()[axis]}
infer_meta :
func : UnbindInferMeta
kernel :
func : unbind
backward : unbind_grad
# unfold
- api : unfold
args : (Tensor x, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations)
output : Tensor
infer_meta :
func : UnfoldInferMeta
kernel :
func : unfold
backward : unfold_grad
- api : uniform_random
args : (IntArray shape, DataType dtype, float min, float max, int seed, Place place={})
output : Tensor(out)
infer_meta :
func : UniformRandomInferMeta
param: [shape, dtype, min, max, seed]
kernel :
func : uniform_random
param: [shape, dtype, min, max, seed]
data_type : dtype
backend : place
# The `axis` argument of Python API paddle.unique is not vector
- api : unique
args : (Tensor x, bool return_index, bool return_inverse, bool return_counts, int[] axis, DataType dtype=DataType::INT64)
output : Tensor(out), Tensor(indices), Tensor(inverse), Tensor(counts)
infer_meta :
func : UniqueInferMeta
kernel :
func : unique
data_type : x
- api : unsqueeze
args : (Tensor x, IntArray axis)
output : Tensor(out), Tensor(xshape)
infer_meta :
func : UnsqueezeInferMeta
kernel :
func : unsqueeze
view: (x -> out)
intermediate : xshape
backward : unsqueeze_grad
# viterbi_decode
- api : viterbi_decode
args : (Tensor input, Tensor transition, Tensor length, bool include_bos_eos_tag)
output : Tensor(scores), Tensor(path)
infer_meta :
func : ViterbiDecodeInferMeta
kernel :
func : viterbi_decode
data_type : input
- api : where
args : (Tensor condition, Tensor x, Tensor y)
output : Tensor
infer_meta :
func : WhereInferMeta
kernel :
func : where
backward : where_grad
# where_index
- api : where_index
args : (Tensor condition)
output : Tensor
infer_meta :
func : WhereIndexInferMeta
kernel :
func : where_index
# yolo_box
- api : yolo_box
args : (Tensor x, Tensor img_size, int[] anchors, int class_num, float conf_thresh, int downsample_ratio, bool clip_bbox, float scale_x_y=1.0, bool iou_aware=false, float iou_aware_factor=0.5)
output : Tensor(boxes), Tensor(scores)
infer_meta :
func : YoloBoxInferMeta
kernel :
func : yolo_box
data_type : x
- api : zeros_like
args : (Tensor x, DataType dtype=DataType::UNDEFINED, Place place = {})
output : Tensor
invoke : full_like(x, 0, dtype, place)
python/paddle/utils/code_gen/backward.yaml
浏览文件 @ fcd32950
#- backward_api : einsum_grad
#forward : einsum (Tensor[] x, str equation) -> Tensor(out), Tensor[](inner_cache)
#args : (Tensor[] x, Tensor[] inner_cache, Tensor out_grad, str equation)
#output : Tensor[](x_grad){x.size()}
#infer_meta :
#func : UnchangedMultiInferMeta
#param : [x]
#kernel :
#func : einsum_grad
- backward_api : abs_double_grad
forward : abs_grad (Tensor x, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_x_grad)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : abs_double_grad
data_transform:
skip_transform : grad_x_grad
- backward_api : abs_grad
forward : abs (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : abs_grad
data_transform:
skip_transform : out_grad
backward : abs_double_grad
- backward_api : acos_grad
forward : acos (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : acos_grad
inplace : (out_grad -> x_grad)
- backward_api : acosh_grad
forward : acosh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : acosh_grad
inplace : (out_grad -> x_grad)
- backward_api : add_double_grad
forward : add_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [grad_out]
kernel :
func : add_double_grad
optional : grad_x_grad, grad_y_grad
backward : add_triple_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : add_grad
forward : add (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : add_grad
no_need_buffer : x, y
backward : add_double_grad
inplace : (out_grad -> x_grad)
- backward_api : add_n_grad
forward : add_n (Tensor[] x) -> Tensor(out)
args : (Tensor[] x, Tensor out_grad)
output : Tensor[](x_grad){x.size()}
invoke : add_n_grad_impl(x, out_grad, x_grad)
no_need_buffer : x
- backward_api : add_triple_grad
forward : add_double_grad (Tensor y, Tensor grad_out, Tensor grad_grad_x, Tensor grad_grad_y, int axis = -1) -> Tensor(grad_grad_out)
args : (Tensor grad_grad_x, Tensor grad_grad_y, Tensor grad_grad_out_grad, int axis = -1)
output : Tensor(grad_grad_x_grad), Tensor(grad_grad_y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [grad_grad_x, grad_grad_y]
kernel :
func : add_triple_grad
inplace : (grad_grad_out_grad -> grad_grad_x_grad)
- backward_api : addmm_grad
forward : addmm (Tensor input, Tensor x, Tensor y, float alpha, float beta) -> Tensor(out)
args : (Tensor input, Tensor x, Tensor y, Tensor out_grad, float alpha, float beta)
output : Tensor(input_grad), Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [input, x, y]
kernel :
func : addmm_grad
- backward_api : argsort_grad
forward : argsort (Tensor x, int axis, bool descending) -> Tensor(out), Tensor(indices)
args : (Tensor indices, Tensor x, Tensor out_grad, int axis, bool descending)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : argsort_grad
no_need_buffer : x
- backward_api : asin_grad
forward : asin (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : asin_grad
inplace : (out_grad -> x_grad)
- backward_api : asinh_grad
forward : asinh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : asinh_grad
inplace : (out_grad -> x_grad)
- backward_api : assign_grad
forward : assign (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
kernel :
func : assign
inplace : (out_grad -> x_grad)
- backward_api : assign_out__grad
forward : assign_out_ (Tensor x, Tensor output) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
kernel :
func : assign
inplace : (out_grad -> x_grad)
- backward_api : atan2_grad
forward : atan2 (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : atan2_grad
- backward_api : atan_grad
forward : atan (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : atan_grad
inplace : (out_grad -> x_grad)
- backward_api : atanh_grad
forward : atanh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : atanh_grad
inplace : (out_grad -> x_grad)
- backward_api : batch_norm_double_grad
forward : batch_norm_grad (Tensor x, Tensor scale, Tensor bias, Tensor out_mean, Tensor out_variance, Tensor saved_mean, Tensor saved_variance, Tensor reserve_space, Tensor grad_out, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu) -> Tensor(grad_x), Tensor(grad_scale), Tensor(grad_bias)
args : (Tensor x, Tensor scale, Tensor out_mean, Tensor out_variance, Tensor saved_mean, Tensor saved_variance, Tensor grad_out, Tensor grad_x_grad, Tensor grad_scale_grad, Tensor grad_bias_grad, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [x, scale, x]
kernel :
func : batch_norm_grad_grad
data_type : x
optional : out_mean, out_variance
inplace : (grad_out -> grad_out_grad)
- backward_api : batch_norm_grad
forward : batch_norm (Tensor x, Tensor scale, Tensor bias, Tensor mean, Tensor variance, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu) -> Tensor(out), Tensor(mean_out), Tensor(variance_out), Tensor(saved_mean), Tensor(saved_variance), Tensor(reserve_space)
args : (Tensor x, Tensor scale, Tensor bias, Tensor mean_out, Tensor variance_out, Tensor saved_mean, Tensor saved_variance, Tensor reserve_space, Tensor out_grad, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [x, scale, bias]
kernel :
func : batch_norm_grad
data_type : out_grad
optional : mean_out, variance_out, reserve_space
backward : batch_norm_double_grad
- backward_api : bce_loss_grad
forward : bce_loss (Tensor input, Tensor label) -> Tensor(out)
args : (Tensor input, Tensor label, Tensor out_grad)
output : Tensor(input_grad)
infer_meta :
func : UnchangedInferMeta
param : [input]
kernel :
func : bce_loss_grad
inplace : (out_grad -> input_grad)
- backward_api : brelu_grad
forward : brelu (Tensor x, float t_min, float t_max) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float t_min, float t_max)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : brelu_grad
inplace : (out_grad -> x_grad)
- backward_api : cast_grad
forward : cast (Tensor x, DataType out_dtype) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : cast_grad
data_type : out_grad
no_need_buffer : x
- backward_api : ceil_grad
forward : ceil(Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : ceil_grad
inplace : (out_grad -> x_grad)
- backward_api : celu_double_grad
forward : celu_grad(Tensor x, Tensor grad_out, float alpha) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_out, Tensor grad_x_grad, float alpha)
output : Tensor(x_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, x]
kernel :
func : celu_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : celu_grad
forward : celu(Tensor x, float alpha) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float alpha)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : celu_grad
backward : celu_double_grad
inplace : (out_grad -> x_grad)
- backward_api : cholesky_grad
forward : cholesky (Tensor x, bool upper) -> Tensor(out)
args : (Tensor out, Tensor out_grad, bool upper)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : cholesky_grad
- backward_api : cholesky_solve_grad
forward : cholesky_solve (Tensor x, Tensor y, bool upper) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, bool upper)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : cholesky_solve_grad
- backward_api : clip_double_grad
forward : clip_grad (Tensor x, Tensor grad_out, Scalar min = 0., Scalar max = 0.) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_x_grad, Scalar min = 0., Scalar max = 0.)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : clip_grad
- backward_api : clip_grad
forward : clip (Tensor x, Scalar min, Scalar max) -> Tensor(out)
args : (Tensor x, Tensor out_grad, Scalar min = 0., Scalar max = 0.)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : clip_grad
backward : clip_double_grad
inplace : (out_grad -> x_grad)
- backward_api : concat_double_grad
forward : concat_grad (Tensor[] x, Tensor grad_out, Scalar axis) -> Tensor[](grad_x)
args : (Tensor[] grad_x_grad, Scalar axis = 0)
output : Tensor(grad_out_grad)
infer_meta :
func : ConcatInferMeta
param : [grad_x_grad, axis]
kernel :
func : concat
- backward_api : concat_grad
forward : concat (Tensor[] x, Scalar axis) -> Tensor(out)
args : (Tensor[] x, Tensor out_grad, Scalar axis = 0)
output : Tensor[](x_grad){x.size()}
infer_meta :
func : UnchangedMultiInferMeta
param : [x]
kernel :
func : concat_grad
no_need_buffer : x
backward : concat_double_grad
- backward_api : conj_grad
forward : conj (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : conj
- backward_api : conv2d_grad
forward : conv2d (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(out)
args : (Tensor input, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor(input_grad), Tensor(filter_grad)
invoke : conv2d_grad_impl(input, filter, out_grad, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search, input_grad, filter_grad)
backward : conv2d_grad_grad
- backward_api : conv2d_grad_grad
forward : conv2d_grad (Tensor input, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(grad_input), Tensor(grad_filter)
args : (Tensor input, Tensor filter, Tensor grad_out, Tensor grad_input_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor(input_grad), Tensor(filter_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param: [input, filter, grad_out]
kernel :
func : conv2d_grad_grad
use_gpudnn : true
optional : grad_input_grad, grad_filter_grad
- backward_api : conv2d_transpose_double_grad
forward : conv2d_transpose_grad(Tensor x, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(grad_x), Tensor(grad_filter)
args : (Tensor x, Tensor filter, Tensor grad_out, Tensor grad_x_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(x_grad), Tensor(filter_grad), Tensor(grad_out_grad)
infer_meta :
func : Conv2dTransposeDoubleGradInferMeta
kernel :
func : conv2d_transpose_grad_grad
use_gpudnn : true
- backward_api : conv2d_transpose_grad
forward : conv2d_transpose(Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(out)
args : (Tensor x, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(x_grad), Tensor(filter_grad)
infer_meta :
func : ConvTransposeGradInferMeta
kernel :
func : conv2d_transpose_grad
use_gpudnn : true
backward : conv2d_transpose_double_grad
- backward_api : conv3d_grad
forward : conv3d (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(out)
args : (Tensor input, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor(input_grad), Tensor(filter_grad)
invoke : conv3d_grad_impl(input, filter, out_grad, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search, input_grad, filter_grad)
backward : conv3d_grad_grad
- backward_api : conv3d_grad_grad
forward : conv3d_grad (Tensor input, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(grad_input), Tensor(grad_filter)
args : (Tensor input, Tensor filter, Tensor grad_out, Tensor grad_input_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor(input_grad), Tensor(filter_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param: [input, filter, grad_out]
kernel :
func : conv3d_grad_grad
use_gpudnn : true
optional : grad_input_grad, grad_filter_grad
- backward_api : conv3d_transpose_grad
forward : conv3d_transpose(Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(out)
args : (Tensor x, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(x_grad), Tensor(filter_grad)
infer_meta :
func : ConvTransposeGradInferMeta
kernel :
func : conv3d_transpose_grad
use_gpudnn : true
- backward_api : cos_grad
forward : cos (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : cos_grad
inplace : (out_grad -> x_grad)
- backward_api : cosh_grad
forward : cosh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : cosh_grad
inplace : (out_grad -> x_grad)
- backward_api : cross_entropy_with_softmax_grad
forward : cross_entropy_with_softmax (Tensor input, Tensor label, bool soft_label, bool use_softmax, bool numeric_stable_mode, int ignore_index, int axis) -> Tensor(softmax), Tensor(loss)
args : (Tensor label, Tensor softmax, Tensor loss_grad, bool soft_label, bool use_softmax, bool numeric_stable_mode, int ignore_index, int axis)
output : Tensor(input_grad)
infer_meta :
func : CrossEntropyWithSoftmaxGradInferMeta
kernel :
func : cross_entropy_with_softmax_grad
data_type : softmax
inplace : (softmax -> input_grad)
- backward_api : cross_grad
forward : cross (Tensor x, Tensor y, int axis = 9) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : cross_grad
- backward_api : cumprod_grad
forward : cumprod (Tensor x, int dim) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int dim)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : cumprod_grad
- backward_api : cumsum_grad
forward : cumsum(Tensor x, int axis, bool flatten, bool exclusive, bool reverse) -> Tensor(out)
infer_meta :
func : UnchangedInferMeta
param : [x]
args : (Tensor out_grad, int axis, bool flatten, bool exclusive, bool reverse)
output : Tensor(x_grad)
invoke : cumsum(out_grad, axis, flatten, exclusive, !reverse)
- backward_api : deformable_conv_grad
forward : deformable_conv(Tensor x, Tensor offset, Tensor filter, Tensor mask, int[] strides, int[] paddings, int[] dilations, int deformable_groups, int groups, int im2col_step) -> Tensor(out)
args : (Tensor x, Tensor offset, Tensor filter, Tensor mask, Tensor out_grad, int[] strides, int[] paddings, int[] dilations, int deformable_groups, int groups, int im2col_step)
output : Tensor(x_grad), Tensor(offset_grad), Tensor(filter_grad), Tensor(mask_grad)
infer_meta :
func : DeformableConvGradInferMeta
kernel :
func : deformable_conv_grad
data_type : x
optional : mask
- backward_api : depthwise_conv2d_grad
forward : depthwise_conv2d (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn) -> Tensor(out)
args : (Tensor input, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn)
output : Tensor(input_grad), Tensor(filter_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [input, filter]
kernel :
func : depthwise_conv2d_grad
param : [input, filter, out_grad, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search, fuse_relu]
use_gpudnn : use_gpudnn
backward : depthwise_conv2d_grad_grad
- backward_api : depthwise_conv2d_grad_grad
forward : depthwise_conv2d_grad (Tensor input, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn) -> Tensor(grad_input), Tensor(grad_filter)
args : (Tensor input, Tensor filter, Tensor grad_out, Tensor grad_input_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu)
output : Tensor(input_grad), Tensor(filter_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param: [input, filter, grad_out]
kernel :
func : depthwise_conv2d_grad_grad
optional : grad_input_grad, grad_filter_grad
- backward_api : depthwise_conv2d_transpose_grad
forward : depthwise_conv2d_transpose(Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(out)
args : (Tensor x, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(x_grad), Tensor(filter_grad)
infer_meta :
func : ConvTransposeGradInferMeta
kernel :
func : depthwise_conv2d_transpose_grad
- backward_api : det_grad
forward : det (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : determinant_grad
- backward_api : diagonal_grad
forward : diagonal (Tensor x, int offset, int axis1, int axis2) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int offset = 0, int axis1 = 0, int axis2 = 1)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : diagonal_grad
no_need_buffer : x
- backward_api : digamma_grad
forward : digamma (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : digamma_grad
- backward_api : dist_grad
forward : dist (Tensor x, Tensor y, float p) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, float p)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : dist_grad
- backward_api : divide_double_grad
forward : divide_grad (Tensor x, Tensor y, Tensor out, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor y, Tensor out, Tensor grad_x, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
output : Tensor(y_grad), Tensor(out_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [y, grad_x, grad_x]
kernel :
func : divide_double_grad
data_type : out
optional : grad_x_grad, grad_y_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : divide_grad
forward : divide (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : divide_grad
backward : divide_double_grad
- backward_api : dropout_grad
forward : dropout (Tensor x, Tensor seed_tensor, float p, bool is_test, str mode, int seed, bool fix_seed) -> Tensor(out), Tensor(mask)
args : (Tensor mask, Tensor out_grad, float p, bool is_test, str mode)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : dropout_grad
- backward_api : eigh_grad
forward : eigh (Tensor x, str uplo) -> Tensor(out_w), Tensor(out_v)
args : (Tensor out_w, Tensor out_v, Tensor out_w_grad, Tensor out_v_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_v]
kernel :
func : eigh_grad
data_type : out_v
data_transform:
skip_transform : out_w, out_w_grad
- backward_api : einsum_grad
forward : einsum (Tensor[] x, str equation) -> Tensor(out), Tensor[](inner_cache), Tensor[](x_shape)
args : (Tensor[] x_shape, Tensor[] inner_cache, Tensor out_grad, str equation)
output : Tensor[](x_grad){x.size()}
infer_meta :
func : UnchangedMultiInferMeta
param : [x_shape]
kernel :
func : einsum_grad
- backward_api : elementwise_pow_grad
forward : elementwise_pow(Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis=-1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : elementwise_pow_grad
- backward_api : elu_double_grad
forward : elu_grad (Tensor x, Tensor out, Tensor grad_out, float alpha)-> Tensor(grad_x)
args : (Tensor x, Tensor grad_out, Tensor grad_x_grad, float alpha)
output : Tensor(x_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, x]
kernel :
func : elu_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : elu_grad
forward : elu (Tensor x, float alpha) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, float alpha)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : elu_grad
backward : elu_double_grad
inplace : (out_grad -> x_grad)
- backward_api : embedding_grad
forward : embedding (Tensor x, Tensor weight, int64_t padding_idx=-1, bool sparse=false) -> Tensor(out)
args : (Tensor x, Tensor weight, Tensor out_grad, int64_t padding_idx=-1, bool sparse=false)
output : Tensor(weight_grad)
invoke : embedding_grad_impl(x, weight, out_grad, padding_idx, sparse, weight_grad)
- backward_api : erf_grad
forward : erf (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : erf_grad
data_type : out_grad
- backward_api : erfinv_grad
forward : erfinv (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : erfinv_grad
- backward_api : exp_grad
forward : exp (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : exp_grad
inplace : (out_grad -> x_grad)
- backward_api : expand_as_grad
forward : expand_as (Tensor x, Tensor y, int[] target_shape) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int[] target_shape)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : expand_as_grad
no_need_buffer : x
- backward_api : expand_double_grad
forward : expand_grad (Tensor x, Tensor grad_out, IntArray shape) -> Tensor(grad_x)
args : (Tensor grad_x_grad, IntArray shape)
output : Tensor(grad_out_grad)
infer_meta :
func : ExpandInferMeta
kernel :
func : expand
- backward_api : expand_grad
forward : expand (Tensor x, IntArray shape) -> Tensor(out)
args : (Tensor x, Tensor out_grad, IntArray shape)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : expand_grad
no_need_buffer : x
backward : expand_double_grad
- backward_api : expm1_grad
forward : expm1 (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : expm1_grad
inplace : (out_grad -> x_grad)
- backward_api : flatten_grad
forward : flatten(Tensor x, int start_axis, int stop_axis) -> Tensor(out), Tensor(xshape)
args : (Tensor xshape, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : KernelWithXShapeInferMeta
param : [xshape]
kernel :
func : flatten_grad
data_type: out_grad
backend: out_grad
layout: out_grad
inplace : (out_grad -> x_grad)
- backward_api : flip_grad
forward : flip (Tensor x, int[] axis) -> Tensor(out)
args : (Tensor out_grad, int[] axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : flip
- backward_api : floor_grad
forward : floor(Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : floor_grad
inplace : (out_grad -> x_grad)
- backward_api : fmax_grad
forward : fmax(Tensor x, Tensor y, int axis) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : fmax_grad
- backward_api : fmin_grad
forward : fmin(Tensor x, Tensor y, int axis) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : fmin_grad
- backward_api : frobenius_norm_grad
forward : frobenius_norm(Tensor x, int64_t[] axis, bool keep_dim, bool reduce_all) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis, bool keep_dim, bool reduce_all)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : frobenius_norm_grad
- backward_api : gather_grad
forward : gather(Tensor x, Tensor index, Scalar axis=0) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad, Scalar axis=0, bool overwrite=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
data_type: x
func : gather_grad
no_need_buffer : x
- backward_api : gather_nd_grad
forward : gather_nd (Tensor x, Tensor index) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : gather_nd_grad
no_need_buffer : x
- backward_api : gelu_grad
forward : gelu(Tensor x, bool approximate) -> Tensor(out)
args : (Tensor x, Tensor out_grad, bool approximate)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : gelu_grad
- backward_api : graph_send_recv_grad
forward : graph_send_recv (Tensor x, Tensor src_index, Tensor dst_index, str pool_type = "SUM", int64_t out_size = 0) -> Tensor(out), Tensor(dst_count)
args : (Tensor x, Tensor src_index, Tensor dst_index, Tensor out, Tensor dst_count, Tensor out_grad, str pool_type = "SUM")
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param : [x]
kernel :
func : graph_send_recv_grad
data_type : out_grad
optional: out, dst_count
- backward_api : group_norm_grad
forward : group_norm (Tensor x, Tensor scale, Tensor bias, float epsilon, int groups, str data_layout) -> Tensor(y), Tensor(mean), Tensor(variance)
args : (Tensor x, Tensor scale, Tensor bias, Tensor y, Tensor mean, Tensor variance, Tensor y_grad, float epsilon, int groups, str data_layout)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [y, scale, bias]
kernel :
func : group_norm_grad
data_type : y_grad
optional: scale, bias
inplace : (y_grad -> x_grad)
- backward_api : gumbel_softmax_grad
forward : gumbel_softmax (Tensor x, float temperature, bool hard, int axis) -> Tensor(out)
args : (Tensor out, Tensor out_grad, int axis)
output : Tensor(x_grad)
infer_meta :
func : GumbelSoftmaxGradInferMeta
param : [out, out_grad, axis]
kernel :
func : gumbel_softmax_grad
- backward_api : hard_shrink_grad
forward : hard_shrink (Tensor x, float threshold) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float threshold)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_shrink_grad
inplace : (out_grad -> x_grad)
- backward_api : hard_sigmoid_grad
forward : hard_sigmoid (Tensor x, float slope, float offset) -> Tensor(out)
args : (Tensor out, Tensor out_grad, float slope, float offset)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : hard_sigmoid_grad
inplace : (out_grad -> x_grad)
- backward_api : hard_swish_grad
forward : hard_swish (Tensor x, float threshold = 6.0, float scale = 6.0, float offset = 3.0) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float threshold, float scale, float offset)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_swish_grad
inplace : (out_grad -> x_grad)
- backward_api : huber_loss_grad
forward : huber_loss (Tensor input, Tensor label, float delta) -> Tensor(out), Tensor(residual)
args : (Tensor residual, Tensor out_grad, float delta)
output : Tensor(input_grad), Tensor(label_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [residual, residual]
kernel :
func : huber_loss_grad
- backward_api : imag_grad
forward : imag (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
invoke : imag_grad_impl(out_grad, x_grad)
- backward_api : index_sample_grad
forward : index_sample (Tensor x, Tensor index) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : index_sample_grad
data_type : out_grad
no_need_buffer : x
- backward_api : index_select_grad
forward : index_select(Tensor x, Tensor index, int dim) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad, int dim)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : index_select_grad
data_type : x
no_need_buffer : x
- backward_api : instance_norm_double_grad
forward : instance_norm_grad(Tensor x, Tensor fwd_scale, Tensor saved_mean, Tensor saved_variance, Tensor grad_y, float epsilon) -> Tensor(grad_x), Tensor(grad_scale), Tensor(grad_bias)
args : (Tensor x, Tensor fwd_scale, Tensor saved_mean, Tensor saved_variance, Tensor grad_y, Tensor grad_x_grad, Tensor grad_scale_grad, Tensor grad_bias_grad, float epsilon)
output : Tensor(x_grad), Tensor(fwd_scale_grad), Tensor(grad_y_grad)
infer_meta :
func : InstanceNormDoubleGradInferMeta
kernel :
func : instance_norm_double_grad
data_type : x
optional : fwd_scale, grad_x_grad, grad_scale_grad, grad_bias_grad
- backward_api : instance_norm_grad
forward : instance_norm(Tensor x, Tensor scale, Tensor bias, float epsilon) -> Tensor(y), Tensor(saved_mean), Tensor(saved_variance)
args : (Tensor x, Tensor scale, Tensor saved_mean, Tensor saved_variance, Tensor y_grad, float epsilon)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
infer_meta :
func : InstanceNormGradInferMeta
kernel :
func : instance_norm_grad
data_type : x
optional : scale
backward : instance_norm_double_grad
- backward_api : kldiv_loss_grad
forward : kldiv_loss(Tensor x, Tensor label, str reduction) -> Tensor(out)
args : (Tensor x, Tensor label, Tensor out_grad, str reduction)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : kldiv_loss_grad
no_need_buffer : x
- backward_api : kron_grad
forward : kron (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : kron_grad
data_type : out_grad
- backward_api : kthvalue_grad
forward : kthvalue(Tensor x, int k, int axis, bool keepdim) -> Tensor(out), Tensor(indices)
args : (Tensor x, Tensor indices, Tensor out_grad, int k, int axis, bool keepdim)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : kthvalue_grad
- backward_api : label_smooth_grad
forward : label_smooth (Tensor label, Tensor prior_dist, float epsilon) -> Tensor(out)
args : (Tensor out_grad, float epsilon)
output : Tensor(label_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : label_smooth_grad
- backward_api : layer_norm_grad
forward : layer_norm (Tensor x, Tensor scale, Tensor bias, float epsilon, int begin_norm_axis, bool is_test) -> Tensor(out), Tensor(mean), Tensor(variance)
args : (Tensor x, Tensor scale, Tensor bias, Tensor mean, Tensor variance, Tensor out_grad, float epsilon, int begin_norm_axis, bool is_test)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
infer_meta :
func : LayerNormGradInferMeta
param : [x, scale, bias]
kernel :
func : layer_norm_grad
data_type : out_grad
no_need_buffer : bias
optional : scale, bias
- backward_api : leaky_relu_double_grad
forward : leaky_relu_grad (Tensor x, Tensor grad_out, float alpha) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_x_grad, float alpha)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [grad_x_grad]
kernel :
func : leaky_relu_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : leaky_relu_grad
forward : leaky_relu (Tensor x, float alpha) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float alpha)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : leaky_relu_grad
backward : leaky_relu_double_grad
inplace : (out_grad -> x_grad)
- backward_api : lerp_grad
forward : lerp (Tensor x, Tensor y, Tensor weight) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor weight, Tensor out, Tensor out_grad)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : lerp_grad
- backward_api : lgamma_grad
forward : lgamma(Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : lgamma_grad
- backward_api : log10_grad
forward : log10 (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : log10_grad
inplace : (out_grad -> x_grad)
- backward_api : log1p_grad
forward : log1p (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : log1p_grad
inplace : (out_grad -> x_grad)
- backward_api : log2_grad
forward : log2 (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : log2_grad
inplace : (out_grad -> x_grad)
- backward_api : log_double_grad
forward : log_grad (Tensor x, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_out, Tensor grad_x_grad)
output : Tensor(x_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, x]
kernel :
func : log_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : log_grad
forward : log (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : log_grad
backward : log_double_grad
inplace : (out_grad -> x_grad)
- backward_api : log_loss_grad
forward : log_loss (Tensor input, Tensor label, float epsilon) -> Tensor(out)
args : (Tensor input, Tensor label, Tensor out_grad, float epsilon)
output : Tensor(input_grad)
infer_meta :
func : UnchangedInferMeta
param : [input]
kernel :
func : log_loss_grad
- backward_api : log_softmax_grad
forward : log_softmax(Tensor x, int axis) -> Tensor(out)
args : (Tensor out, Tensor out_grad, int axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out]
kernel :
func : log_softmax_grad
- backward_api : logcumsumexp_grad
forward : logcumsumexp(Tensor x, int axis, bool flatten, bool exclusive, bool reverse) -> Tensor(out)
infer_meta :
func : UnchangedInferMeta
param : [x]
args : (Tensor x, Tensor out, Tensor out_grad, int axis, bool flatten, bool exclusive, bool reverse)
output : Tensor(x_grad)
kernel :
func : logcumsumexp_grad
- backward_api : logit_grad
forward : logit (Tensor x, float eps = 1e-6f) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float eps)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : logit_grad
- backward_api : logsigmoid_grad
forward : logsigmoid (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : logsigmoid_grad
inplace : (out_grad -> x_grad)
- backward_api : logsumexp_grad
forward : logsumexp(Tensor x, int64_t[] axis, bool keepdim, bool reduce_all) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis, bool keepdim, bool reduce_all)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : logsumexp_grad
- backward_api : masked_select_grad
forward : masked_select (Tensor x, Tensor mask) -> Tensor(out)
args : (Tensor x, Tensor mask, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : masked_select_grad
data_type : x
no_need_buffer : x
- backward_api : matmul_double_grad
forward : matmul_grad (Tensor x, Tensor y, Tensor grad_out, bool transpose_x=false, bool transpose_y=false) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor x, Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, bool transpose_x=false, bool transpose_y=false)
output : Tensor(x_grad), Tensor(y_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [x, y, grad_out]
kernel :
func : matmul_double_grad
backward : matmul_triple_grad
optional : grad_x_grad, grad_y_grad
- backward_api : matmul_grad
forward : matmul (Tensor x, Tensor y, bool transpose_x=false, bool transpose_y=false) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, bool transpose_x=false, bool transpose_y=false)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : matmul_grad
backward : matmul_double_grad
- backward_api : matmul_triple_grad
forward : matmul_double_grad (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, bool transpose_x=false, bool transpose_y=false) -> Tensor(grad_x), Tensor(grad_y), Tensor(grad_grad_out)
args : (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, Tensor grad_x_grad, Tensor grad_y_grad, Tensor grad_grad_out_grad, bool transpose_x=false, bool transpose_y=false)
output : Tensor(x_grad), Tensor(y_grad), Tensor(fwd_grad_out_grad), Tensor(fwd_grad_grad_x_grad), Tensor(fwd_grad_grad_y_grad)
infer_meta :
func : GeneralQuinaryGradInferMeta
param : [x, y, fwd_grad_out, fwd_grad_grad_x, fwd_grad_grad_y]
kernel :
func : matmul_triple_grad
optional : grad_x_grad, grad_y_grad, grad_grad_out_grad
- backward_api : matrix_power_grad
forward : matrix_power (Tensor x, int n) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int n)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : matrix_power_grad
- backward_api : max_grad
forward: max (Tensor x, int64_t[] dims={}, bool keep_dim=false) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : max_grad
- backward_api : max_pool2d_with_index_grad
forward : max_pool2d_with_index(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive) -> Tensor(out), Tensor(mask)
args : (Tensor x, Tensor mask, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
output : Tensor(x_grad)
infer_meta :
func : MaxPoolWithIndexGradInferMeta
kernel :
func : max_pool2d_with_index_grad
- backward_api : max_pool3d_with_index_grad
forward : max_pool3d_with_index(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive) -> Tensor(out), Tensor(mask)
args : (Tensor x, Tensor mask, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
output : Tensor(x_grad)
infer_meta :
func : MaxPoolWithIndexGradInferMeta
kernel :
func : max_pool3d_with_index_grad
- backward_api : maximum_grad
forward : maximum(Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis=-1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : maximum_grad
- backward_api : maxout_grad
forward : maxout(Tensor x, int groups, int axis) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int groups, int axis)
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param: [x]
kernel :
func : maxout_grad
- backward_api : mean_all_grad
forward : mean_all(Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : mean_all_grad
- backward_api : mean_double_grad
forward: mean_grad (Tensor x, Tensor grad_out, int64_t[] dims={}, bool keep_dim=false, bool reduce_all = false) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(grad_out_grad)
invoke : mean(grad_x_grad, dims, keep_dim)
- backward_api : mean_grad
forward: mean (Tensor x, int64_t[] dims={}, bool keep_dim=false) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : mean_grad
backward : mean_double_grad
no_need_buffer : x
- backward_api : meshgrid_grad
forward : meshgrid (Tensor[] inputs) -> Tensor[](outputs)
args : (Tensor[] inputs, Tensor[] outputs_grad)
output : Tensor[](inputs_grad){inputs.size()}
infer_meta :
func : MeshgridGradInferMeta
kernel :
func : meshgrid_grad
- backward_api : min_grad
forward: min (Tensor x, int64_t[] dims={}, bool keep_dim=false) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : min_grad
- backward_api : minimum_grad
forward : minimum(Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis=-1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : minimum_grad
- backward_api : mish_grad
forward : mish (Tensor x, float threshold) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float threshold)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : mish_grad
inplace : (out_grad -> x_grad)
- backward_api : mode_grad
forward : mode(Tensor x, int axis, bool keepdim) -> Tensor(out), Tensor(indices)
args : (Tensor x, Tensor indices, Tensor out_grad, int axis, bool keepdim)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : mode_grad
- backward_api : modulo_grad
forward : modulo (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : modulo_grad
no_need_buffer : x, y
- backward_api : multi_dot_grad
forward : multi_dot (Tensor[] x) -> Tensor(out)
args : (Tensor[] x, Tensor out_grad)
output : Tensor[](x_grad) {x.size()}
infer_meta :
func : MultiDotGradInferMeta
kernel :
func : multi_dot_grad
- backward_api : multiplex_grad
forward : multiplex (Tensor[] ins, Tensor ids) -> Tensor(out)
args : (Tensor[] ins, Tensor ids, Tensor out_grad)
output : Tensor[](ins_grad){ins.size()}
infer_meta :
func : MultiplexGradInferMeta
param : [ids, out_grad]
kernel :
func : multiplex_grad
param : [ids, out_grad]
- backward_api : multiply_double_grad
forward : multiply_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor x, Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [x, y, grad_out]
kernel :
func : multiply_double_grad
optional : grad_x_grad, grad_y_grad
backward : multiply_triple_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : multiply_grad
forward : multiply (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : multiply_grad
backward : multiply_double_grad
- backward_api : multiply_triple_grad
forward : multiply_double_grad (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, int aixs = -1) -> Tensor(grad_x), Tensor(grad_y), Tensor(grad_grad_out)
args : (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, Tensor grad_x_grad, Tensor grad_y_grad, Tensor grad_grad_out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad), Tensor(fwd_grad_out_grad), Tensor(fwd_grad_grad_x_grad), Tensor(fwd_grad_grad_y_grad)
infer_meta :
func : GeneralQuinaryGradInferMeta
param : [x, y, fwd_grad_out, x, y]
kernel :
func : multiply_triple_grad
optional : fwd_grad_grad_x, fwd_grad_grad_y, grad_grad_out_grad
- backward_api : mv_grad
forward : mv (Tensor x, Tensor vec) -> Tensor(out)
args : (Tensor x, Tensor vec, Tensor out_grad)
output : Tensor(x_grad), Tensor(vec_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, vec]
kernel :
func : mv_grad
- backward_api : nll_loss_grad
forward : nll_loss (Tensor input, Tensor label, Tensor weight, int64_t ignore_index, str reduction) -> Tensor(out), Tensor(total_weight)
args : (Tensor input, Tensor label, Tensor weight, Tensor total_weight, Tensor out_grad, int64_t ignore_index, str reduction)
output : Tensor(input_grad)
infer_meta :
func : NllLossGradInferMeta
kernel :
func : nll_loss_grad
data_type : input
optional : weight
- backward_api : norm_grad
forward : norm (Tensor x, int axis, float epsilon, bool is_test) -> Tensor(out), Tensor(norm)
args : (Tensor x, Tensor norm, Tensor out_grad, int axis, float epsilon, bool is_test)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : norm_grad
- backward_api : p_norm_grad
forward : p_norm(Tensor x, float porder, int axis, float epsilon, bool keepdim, bool asvector=false) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, float porder, int axis, float epsilon, bool keepdim, bool asvector)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : p_norm_grad
- backward_api : pad3d_double_grad
forward : pad3d_grad(Tensor x, Tensor grad_out, IntArray paddings, str mode, float pad_value, str data_format) -> Tensor(grad_x)
args : (Tensor grad_x_grad, IntArray paddings, str mode, float pad_value, str data_format)
output : Tensor(grad_out_grad)
infer_meta :
func : Pad3dInferMeta
kernel :
func : pad3d
- backward_api : pad3d_grad
forward : pad3d(Tensor x, IntArray paddings, str mode, float pad_value, str data_format) -> Tensor(out)
args : (Tensor x, Tensor out_grad, IntArray paddings, str mode, float pad_value, str data_format)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : pad3d_grad
no_need_buffer : x
backward : pad3d_double_grad
- backward_api : pad_double_grad
forward : pad_grad(Tensor x, Tensor grad_out, int[] paddings, float pad_value) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int[] paddings, float pad_value)
output : Tensor(grad_out_grad)
infer_meta :
func : PadInferMeta
kernel :
func : pad
- backward_api : pad_grad
forward : pad(Tensor x, int[] paddings, float pad_value) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int[] paddings, float pad_value)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : pad_grad
param: [out_grad, paddings, pad_value]
no_need_buffer : x
backward : pad_double_grad
- backward_api : pixel_shuffle_grad
forward : pixel_shuffle (Tensor x, int upscale_factor, str data_format) -> Tensor(out)
args : (Tensor out_grad, int upscale_factor, str data_format)
output : Tensor(x_grad)
infer_meta :
func : PixelShuffleGradInferMeta
kernel :
func : pixel_shuffle_grad
- backward_api : poisson_grad
forward : poisson (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : poisson_grad
- backward_api : pool2d_double_grad
forward : pool2d_grad(Tensor x, Tensor out, Tensor grad_out, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(grad_out_grad)
infer_meta :
func : PoolInferMeta
kernel :
func : pool2d_double_grad
use_gpudnn : true
- backward_api : pool2d_grad
forward : pool2d(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(x_grad)
infer_meta :
func : PoolGradInferMeta
kernel :
func : pool2d_grad
use_gpudnn : true
backward : pool2d_double_grad
- backward_api : pool2d_grad_gpudnn_unused
forward : pool2d_gpudnn_unused(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(x_grad)
infer_meta :
func : PoolGradInferMeta
kernel :
func : pool2d_grad
use_gpudnn : false
- backward_api : pool3d_grad
forward : pool3d(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(x_grad)
infer_meta :
func : PoolGradInferMeta
kernel :
func : pool3d_grad
use_gpudnn : true
- backward_api : pow_grad
forward : pow(Tensor x, Scalar s) -> Tensor(out)
args : (Tensor x, Tensor out_grad, Scalar s=-1)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : pow_grad
inplace : (out_grad -> x_grad)
- backward_api : prelu_grad
forward : prelu(Tensor x, Tensor alpha, str data_format, str mode) -> Tensor(out)
args : (Tensor x, Tensor alpha, Tensor out_grad, str data_format, str mode)
output : Tensor(x_grad), Tensor(alpha_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, alpha]
kernel :
func : prelu_grad
- backward_api : psroi_pool_grad
forward : psroi_pool (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, int output_channels, float spatial_scale) -> Tensor(out)
args : (Tensor x, Tensor boxes, Tensor boxes_num, Tensor out_grad, int pooled_height, int pooled_width, int output_channels, float spatial_scale)
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param : [x]
kernel :
func : psroi_pool_grad
data_type : x
optional : boxes_num
# output is optional
- backward_api : put_along_axis_grad
forward : put_along_axis (Tensor x, Tensor index, Tensor value, int axis, str reduce) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad, int axis, str reduce)
output : Tensor(x_grad), Tensor(value_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, index]
kernel :
func : put_along_axis_grad
- backward_api : real_grad
forward : real (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
invoke : real_grad_impl(out_grad, x_grad)
- backward_api : reciprocal_grad
forward : reciprocal (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : reciprocal_grad
inplace : (out_grad -> x_grad)
- backward_api : reduce_prod_grad
forward : reduce_prod (Tensor x, int64_t[] dims, bool keep_dim, bool reduce_all) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] dims, bool keep_dim, bool reduce_all)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : prod_grad
- backward_api : relu_double_grad
forward : relu_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor grad_x_grad)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : relu_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : relu_grad
forward : relu (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : relu_grad
backward: relu_double_grad
inplace : (out_grad -> x_grad)
- backward_api : reshape_double_grad
forward : reshape_grad (Tensor xshape, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor grad_out, Tensor grad_x_grad)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [grad_out]
kernel :
func : reshape_double_grad
no_need_buffer : grad_out
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : reshape_grad
forward : reshape (Tensor x, IntArray shape) -> Tensor(out), Tensor(xshape)
args : (Tensor xshape, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : KernelWithXShapeInferMeta
param : [xshape]
kernel :
func : reshape_grad
param : [out_grad]
data_type: out_grad
backend: out_grad
layout: out_grad
backward : reshape_double_grad
inplace : (out_grad -> x_grad)
- backward_api : roi_align_grad
forward : roi_align (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale, int sampling_ratio, bool aligned) -> Tensor(out)
args : (Tensor x, Tensor boxes, Tensor boxes_num, Tensor out_grad, int pooled_height, int pooled_width, float spatial_scale, int sampling_ratio, bool aligned)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : roi_align_grad
data_type : boxes
no_need_buffer : x
optional : boxes_num
- backward_api : roi_pool_grad
forward : roi_pool (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale) -> Tensor(out), Tensor(arg_max)
args : (Tensor x, Tensor boxes, Tensor boxes_num, Tensor arg_max, Tensor out_grad, int pooled_height, int pooled_width, float spatial_scale)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : roi_pool_grad
data_type : x
optional : boxes_num
- backward_api : roll_grad
forward : roll(Tensor x, IntArray shifts, int64_t[] axis) -> Tensor(out)
args : (Tensor x, Tensor out_grad, IntArray shifts, int64_t[] axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : roll_grad
data_type : x
no_need_buffer : x
- backward_api : round_grad
forward : round(Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : round_grad
inplace : (out_grad -> x_grad)
- backward_api : rsqrt_double_grad
forward : rsqrt_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor grad_x, Tensor grad_x_grad)
output : Tensor(out_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [out, out]
kernel :
func : rsqrt_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : rsqrt_grad
forward : rsqrt (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : rsqrt_grad
backward : rsqrt_double_grad
inplace : (out_grad -> x_grad)
- backward_api : scale_double_grad
forward : scale_grad (Tensor grad_out, Scalar scale, float bias, bool bias_after_scale) -> Tensor(grad_x)
args : (Tensor grad_x_grad, Scalar scale=1.0, float bias=0.0, bool bias_after_scale=true)
output : Tensor(grad_out_grad)
invoke : scale(grad_x_grad, scale, 0.0, bias_after_scale)
backward : scale_triple_grad
- backward_api : scale_grad
forward : scale (Tensor x, Scalar scale, float bias, bool bias_after_scale) -> Tensor(out)
args : (Tensor out_grad, Scalar scale=1.0, float bias=0.0, bool bias_after_scale=true)
output : Tensor(x_grad)
invoke : scale(out_grad, scale, 0.0, bias_after_scale)
backward : scale_double_grad
inplace : (out_grad -> x_grad)
- backward_api : scale_triple_grad
forward : scale_double_grad (Tensor grad_grad_x, Scalar scale, float bias, bool bias_after_scale) -> Tensor(grad_grad_out)
args : (Tensor grad_grad_out_grad, Scalar scale=1.0, float bias=0.0, bool bias_after_scale=true)
output : Tensor(grad_grad_x_grad)
invoke : scale(grad_grad_out_grad, scale, 0.0, bias_after_scale)
- backward_api : scatter_grad
forward : scatter (Tensor x, Tensor index, Tensor updates, bool overwrite) -> Tensor(out)
args : (Tensor index, Tensor updates, Tensor out_grad, bool overwrite)
output : Tensor(x_grad), Tensor(updates_grad)
infer_meta :
func : ScatterGradInferMeta
param : [index, updates, out_grad, overwrite]
kernel :
func : scatter_grad
no_need_buffer : updates
- backward_api : scatter_nd_add_grad
forward : scatter_nd_add (Tensor x, Tensor index, Tensor updates) -> Tensor(out)
args : (Tensor index, Tensor updates, Tensor out_grad)
output : Tensor(x_grad), Tensor(updates_grad)
infer_meta :
func : ScatterNdAddGradInferMeta
param : [index, updates, out_grad]
kernel :
func : scatter_nd_add_grad
no_need_buffer : updates
- backward_api : segment_pool_grad
forward : segment_pool (Tensor x, Tensor segment_ids, str pooltype) -> Tensor(out), Tensor(summed_ids)
args : (Tensor x, Tensor segment_ids, Tensor out, Tensor summed_ids, Tensor out_grad, str pooltype)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : segment_pool_grad
data_type : x
optional : summed_ids
- backward_api : selu_grad
forward : selu (Tensor x, float scale, float alpha) -> Tensor(out)
args : (Tensor out, Tensor out_grad, float scale, float alpha)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : selu_grad
- backward_api : sigmoid_cross_entropy_with_logits_grad
forward : sigmoid_cross_entropy_with_logits (Tensor x, Tensor label, bool normalize, int ignore_index) -> Tensor(out)
args : (Tensor x, Tensor label, Tensor out_grad, bool normalize, int ignore_index)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : sigmoid_cross_entropy_with_logits_grad
inplace : (out_grad -> x_grad)
- backward_api : sigmoid_double_grad
forward : sigmoid_grad (Tensor out, Tensor fwd_grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor fwd_grad_out, Tensor grad_x_grad)
output : Tensor(out_grad), Tensor(fwd_grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [out, fwd_grad_out]
kernel :
func : sigmoid_double_grad
backward : sigmoid_triple_grad
inplace : (grad_x_grad -> fwd_grad_out_grad)
- backward_api : sigmoid_grad
forward : sigmoid (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : sigmoid_grad
backward : sigmoid_double_grad
inplace : (out_grad -> x_grad)
- backward_api : sigmoid_triple_grad
forward : sigmoid_double_grad (Tensor out, Tensor fwd_grad_out, Tensor grad_grad_x) -> Tensor(grad_out), Tensor(grad_grad_out)
args : (Tensor out, Tensor fwd_grad_out, Tensor grad_grad_x, Tensor grad_out_grad, Tensor grad_grad_out_grad)
output : Tensor(out_grad), Tensor(fwd_grad_out_grad), Tensor(grad_grad_x_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [out, fwd_grad_out, grad_grad_x]
kernel :
func : sigmoid_triple_grad
optional : grad_grad_out_grad
inplace : (grad_grad_x -> fwd_grad_out_grad)
- backward_api : silu_grad
forward : silu (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : silu_grad
inplace : (out_grad -> x_grad)
- backward_api : sin_grad
forward : sin (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : sin_grad
inplace : (out_grad -> x_grad)
- backward_api : sinh_grad
forward : sinh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : sinh_grad
inplace : (out_grad -> x_grad)
- backward_api : slice_grad
forward : slice (Tensor input, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis) -> Tensor(out)
args : (Tensor input, Tensor out_grad, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis)
output : Tensor(input_grad)
infer_meta :
func : UnchangedInferMeta
param : [input]
kernel :
func : slice_grad
no_need_buffer : input
- backward_api : soft_shrink_grad
forward : soft_shrink (Tensor x, float lambda) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float lambda)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : soft_shrink_grad
inplace : (out_grad -> x_grad)
- backward_api : softmax_grad
forward : softmax (Tensor x, int axis) -> Tensor(out)
args : (Tensor out, Tensor out_grad, int axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : softmax_grad
use_gpudnn : true
- backward_api : split_grad
forward : split (Tensor x, IntArray num_or_sections, Scalar axis) -> Tensor[](out)
args : (Tensor[] out_grad, Scalar axis = -1)
output : Tensor(x_grad)
invoke : concat( out_grad, axis)
# TODO(zhangyunfei) The config of double grad and triple grad will be supported in the future.
- backward_api : sqrt_double_grad
forward : sqrt_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor grad_x, Tensor grad_x_grad)
output : Tensor(out_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [out, out]
kernel :
func : sqrt_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : sqrt_grad
forward : sqrt (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : sqrt_grad
backward : sqrt_double_grad
inplace : (out_grad -> x_grad)
- backward_api : square_double_grad
forward : square_grad (Tensor x, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_out, Tensor grad_x_grad)
output : Tensor(x_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, x]
kernel :
func : square_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : square_grad
forward : square (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : square_grad
backward : square_double_grad
inplace : (out_grad -> x_grad)
- backward_api : squeeze_double_grad
forward : squeeze_grad(Tensor xshape, Tensor grad_out, int[] axes) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int[] axes)
output : Tensor(grad_out_grad)
invoke: squeeze(grad_x_grad, axes)
- backward_api : squeeze_grad
forward : squeeze(Tensor x, int[] axes) -> Tensor(out), Tensor(xshape)
args : (Tensor xshape, Tensor out_grad, int[] axes)
output : Tensor(x_grad)
infer_meta :
func : KernelWithXShapeInferMeta
param: [xshape]
kernel :
func : squeeze_grad
inplace : (out_grad -> x_grad)
backward: squeeze_double_grad
- backward_api : stack_grad
forward : stack (Tensor[] x, int axis) -> Tensor(out)
args : (Tensor[] x, Tensor out_grad, int axis)
output : Tensor[](x_grad){x.size()}
infer_meta :
func : StackGradInferMeta
param: [out_grad, axis]
kernel :
func : stack_grad
param : [out_grad, axis]
no_need_buffer : x
- backward_api : strided_slice_grad
forward : strided_slice (Tensor x, int[] axes, IntArray starts, IntArray ends, IntArray strides) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int[] axes, IntArray starts, IntArray ends, IntArray strides)
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param : [x]
kernel :
func : strided_slice_grad
no_need_buffer : x
- backward_api : subtract_double_grad
forward : subtract_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [grad_out]
kernel :
func : subtract_double_grad
optional : grad_x_grad, grad_y_grad
no_need_buffer : y, grad_out
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : subtract_grad
forward : subtract (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : subtract_grad
no_need_buffer : x, y
backward : subtract_double_grad
inplace : (out_grad -> x_grad)
- backward_api : sum_double_grad
forward : sum_grad (Tensor x, Tensor grad_out, int64_t[] dims, bool keep_dim, bool reduce_all=false) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(grad_out_grad)
invoke : sum(grad_x_grad, dims, grad_x_grad.dtype(), keep_dim)
backward : sum_triple_grad
- backward_api : sum_grad
forward : sum (Tensor x, int64_t[] dims={}, DataType out_dtype=DataType::UNDEFINED, bool keep_dim=false) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int64_t[] dims, bool keep_dim, bool reduce_all=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : sum_grad
no_need_buffer : x
backward : sum_double_grad
- backward_api : sum_triple_grad
forward : sum_double_grad (Tensor grad_grad_x, int64_t[] dims={}, bool keep_dim=false) -> Tensor(grad_grad_out)
args : (Tensor grad_grad_x, Tensor grad_grad_out_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(grad_grad_x_grad)
invoke : sum_grad(grad_grad_x, grad_grad_out_grad, dims, keep_dim, reduce_all, grad_grad_x_grad)
- backward_api : swish_grad
forward : swish (Tensor x, float beta=1.0) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float bete=1.0)
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param : [x]
kernel :
func : swish_grad
inplace : (out_grad -> x_grad)
- backward_api : take_along_axis_grad
forward : take_along_axis (Tensor x, Tensor index, int axis) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad, int axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : take_along_axis_grad
- backward_api : tan_grad
forward : tan (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : tan_grad
inplace : (out_grad -> x_grad)
- backward_api : tanh_double_grad
forward : tanh_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor grad_out, Tensor grad_x_grad)
output : Tensor(out_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [out, out]
kernel :
func : tanh_double_grad
backward : tanh_triple_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : tanh_grad
forward : tanh (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : tanh_grad
backward : tanh_double_grad
inplace : (out_grad -> x_grad)
- backward_api : tanh_shrink_grad
forward : tanh_shrink (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : tanh_shrink_grad
inplace : (out_grad -> x_grad)
- backward_api : tanh_triple_grad
forward : tanh_double_grad (Tensor out, Tensor grad_out_forward, Tensor grad_x_grad_forward) -> Tensor(grad_out_new), Tensor(grad_out_grad)
args : (Tensor out, Tensor grad_out_forward, Tensor grad_x_grad_forward, Tensor grad_out_new_grad, Tensor grad_out_grad_grad)
output : Tensor(out_grad), Tensor(grad_out_forward_grad), Tensor(grad_x_grad_forward_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [out, out, grad_x_grad_forward]
kernel :
func : tanh_triple_grad
inplace : (grad_x_grad_forward -> grad_out_forward_grad)
- backward_api : thresholded_relu_grad
forward : thresholded_relu (Tensor x, float threshold) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float threshold)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : thresholded_relu_grad
inplace : (out_grad -> x_grad)
- backward_api : tile_double_grad
forward : tile_grad (Tensor x, Tensor grad_out, IntArray repeat_times) -> Tensor(grad_x)
args : (Tensor grad_x_grad, IntArray repeat_times)
output : Tensor(grad_out_grad)
infer_meta :
func : TileInferMeta
kernel :
func : tile
- backward_api : tile_grad
forward : tile (Tensor x, IntArray repeat_times) -> Tensor(out)
args : (Tensor x, Tensor out_grad, IntArray repeat_times)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : tile_grad
no_need_buffer : x
backward : tile_double_grad
- backward_api : top_k_grad
forward : top_k (Tensor x, Scalar k, int axis = -1, bool largest = true, bool sorted = true) -> Tensor(out), Tensor(indices)
args : (Tensor x, Tensor indices, Tensor out_grad, Scalar k = -1, int axis = -1, bool largest = true, bool sorted = true)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : top_k_grad
- backward_api : trace_grad
forward : trace (Tensor x, int offset, int axis1, int axis2) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int offset, int axis1, int axis2)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : trace_grad
no_need_buffer : x
- backward_api : transpose_double_grad
forward : transpose_grad (Tensor grad_out, int[] axis) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int[] axis)
output : Tensor(grad_out_grad)
invoke : transpose(grad_x_grad, axis)
- backward_api : transpose_grad
forward : transpose (Tensor x, int[] axis) -> Tensor(out)
args : (Tensor out_grad, int[] axis)
output : Tensor(x_grad)
infer_meta :
func : TransposeGradInferMeta
param : [out_grad, axis]
kernel :
func : transpose_grad
backward : transpose_double_grad
- backward_api : triangular_solve_grad
forward : triangular_solve (Tensor x, Tensor y, bool upper, bool tranpose, bool unitriangular) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, bool upper, bool tranpose, bool unitriangular)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : triangular_solve_grad
- backward_api : tril_triu_grad
forward : tril_triu(Tensor x, int diagonal, bool lower) -> Tensor(out)
args : (Tensor out_grad, int diagonal, bool lower)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : tril_triu_grad
- backward_api : trunc_grad
forward : trunc (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : trunc_grad
- backward_api : unbind_grad
forward : unbind (Tensor input, int axis) -> Tensor[](out)
args : (Tensor[] out_grad, int axis)
output : Tensor(input_grad)
invoke : stack(out_grad, axis)
- backward_api : unfold_grad
forward : unfold (Tensor x, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : unfold_grad
no_need_buffer : x
- backward_api : unsqueeze_double_grad
forward : unsqueeze_grad(Tensor xshape, Tensor grad_out, IntArray axes) -> Tensor(grad_x)
args : (Tensor grad_x_grad, IntArray axes)
output : Tensor(grad_out_grad)
invoke : unsqueeze(grad_x_grad, axes)
- backward_api : unsqueeze_grad
forward : unsqueeze(Tensor x, IntArray axes) -> Tensor(out), Tensor(xshape)
args : (Tensor xshape, Tensor out_grad, IntArray axes)
output : Tensor(x_grad)
infer_meta :
func : KernelWithXShapeInferMeta
param: [xshape]
kernel :
func : unsqueeze_grad
param: [xshape, out_grad]
inplace : (out_grad -> x_grad)
backward : unsqueeze_double_grad
- backward_api : where_grad
forward : where (Tensor condition, Tensor x, Tensor y) -> Tensor(out)
args : (Tensor condition, Tensor x, Tensor y, Tensor out_grad)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : where_grad
no_need_buffer : x, y
python/paddle/utils/code_gen/cross_validate.py
浏览文件 @ fcd32950
...@@ -43,7 +43,7 @@ if __name__ == "__main__": ...@@ -43,7 +43,7 @@ if __name__ == "__main__":
parser.add_argument('--backward_yaml_paths', parser.add_argument('--backward_yaml_paths',
type=str, type=str,
nargs='+', nargs='+',
default=str(current_dir / "backward.yaml.yaml"), default=str(current_dir / "backward_api.parsed.yaml"),
help="backward api yaml file.") help="backward api yaml file.")
args = parser.parse_args() args = parser.parse_args()
......
python/paddle/utils/code_gen/legacy_api.yaml 0 → 100644
浏览文件 @ fcd32950
- api : abs
args : (Tensor x)
output : Tensor
infer_meta :
func : RealAndImagInferMeta
kernel :
func : abs
backward : abs_grad
# accuracy
- api : accuracy
args : (Tensor x, Tensor indices, Tensor label)
output : Tensor(accuracy), Tensor(correct), Tensor(total)
infer_meta :
func : AccuracyInferMeta
kernel :
func : accuracy
dtype : x
# acos
- api : acos
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : acos
backward : acos_grad
# acosh
- api : acosh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : acosh
backward : acosh_grad
- api : adadelta
args : (Tensor param, Tensor grad, Tensor avg_squared_grad, Tensor avg_squared_update, float rho, float epsilon)
output : Tensor(param_out), Tensor(moment_out), Tensor(inf_norm_out)
infer_meta :
func : AdadeltaInferMeta
kernel :
func : adadelta
- api : adam
args : (Tensor param, Tensor grad, Tensor learning_rate, Tensor moment1, Tensor moment2, Tensor beta1_pow, Tensor beta2_pow, Tensor master_param, Tensor skip_update, Scalar beta1, Scalar beta2, Scalar epsilon, bool lazy_mode, int64_t min_row_size_to_use_multithread, bool multi_precision, bool use_global_beta_pow)
output : Tensor(param_out), Tensor(moment1_out), Tensor(moment2_out), Tensor(beta1_pow_out), Tensor(beta2_pow_out), Tensor(master_param_outs)
optional : master_param, skip_update
invoke : adam_impl(param, grad, learning_rate, moment1, moment2, beta1_pow, beta2_pow, master_param, skip_update, beta1, beta2, epsilon, lazy_mode, min_row_size_to_use_multithread, multi_precision, use_global_beta_pow)
- api : adamax
args : (Tensor param, Tensor grad, Tensor learning_rate, Tensor moment, Tensor inf_norm, Tensor beta1_pow, float beta1, float beta2, float epsilon)
output : Tensor(param_out), Tensor(avg_squared_grad_out), Tensor(avg_squared_update_out)
infer_meta :
func : AdamaxInferMeta
kernel :
func : adamax
- api : adamw
args : (Tensor param, Tensor grad, Tensor learning_rate, Tensor moment1, Tensor moment2, Tensor beta1_pow, Tensor beta2_pow, Tensor master_param, Tensor skip_update, Scalar beta1, Scalar beta2, Scalar epsilon, float lr_ratio, float coeff, bool with_decay, bool lazy_mode, int64_t min_row_size_to_use_multithread, bool multi_precision, bool use_global_beta_pow)
output : Tensor(param_out), Tensor(moment1_out), Tensor(moment2_out), Tensor(beta1_pow_out), Tensor(beta2_pow_out), Tensor(master_param_outs)
optional : master_param, skip_update
invoke : adamw_impl(param, grad, learning_rate, moment1, moment2, beta1_pow, beta2_pow, master_param, skip_update, beta1, beta2, epsilon, lr_ratio, coeff, with_decay, lazy_mode, min_row_size_to_use_multithread, multi_precision, use_global_beta_pow)
- api : add
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : add
backward : add_grad
- api : add_n
args : (Tensor[] x)
output : Tensor
infer_meta :
func : AddNInferMeta
kernel :
func : add_n
backward : add_n_grad
- api : addmm
args : (Tensor input, Tensor x, Tensor y, float alpha, float beta)
output : Tensor
infer_meta :
func : AddmmInferMeta
kernel :
func : addmm
backward : addmm_grad
- api : all
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : all
- api : allclose
args : (Tensor x, Tensor y, Scalar rtol, Scalar atol, bool equal_nan)
output : Tensor(out)
infer_meta :
func : AllValueCompareInferMeta
param: [x, y]
kernel :
func : allclose
- api : any
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : any
- api : arange
args : (Tensor start, Tensor end, Tensor step, DataType dtype, Place place={})
output : Tensor
infer_meta :
func : ArangeInferMeta
param : [start, end, step]
kernel :
func : arange
param : [start, end, step]
data_type : dtype
backend : place
data_transform :
support_trans_dtype : start, end, step
# arg_max
- api : argmax
args : (Tensor x, int64_t axis, bool keepdims, bool flatten, int dtype)
output : Tensor
infer_meta :
func : ArgMinMaxInferMeta
kernel :
func : arg_max
# arg_min
- api : argmin
args : (Tensor x, int64_t axis, bool keepdims, bool flatten, int dtype)
output : Tensor
infer_meta :
func : ArgMinMaxInferMeta
kernel :
func : arg_min
# argsort
- api : argsort
args : (Tensor x, int axis, bool descending)
output : Tensor(out), Tensor(indices)
infer_meta :
func : ArgsortInferMeta
kernel :
func : argsort
backward : argsort_grad
# asin
- api : asin
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : asin
backward : asin_grad
# asinh
- api : asinh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : asinh
backward : asinh_grad
# assign
- api : assign
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : assign
backward : assign_grad
- api : assign_out_
args : (Tensor x, Tensor output)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : assign
param : [x]
inplace : (output -> out)
backward : assign_out__grad
# atan
- api : atan
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : atan
backward : atan_grad
- api : atan2
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : Atan2InferMeta
kernel :
func : atan2
backward : atan2_grad
# atanh
- api : atanh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : atanh
backward : atanh_grad
# auc
- api : auc
args : (Tensor x, Tensor label, Tensor stat_pos, Tensor stat_neg, str curve, int num_thresholds, int slide_steps)
output : Tensor(auc), Tensor(stat_pos_out), Tensor(stat_neg_out)
infer_meta :
func : AucInferMeta
kernel :
func : auc
# batch_norm
- api : batch_norm
args : (Tensor x, Tensor scale, Tensor bias, Tensor mean, Tensor variance, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu)
output : Tensor(out), Tensor(mean_out), Tensor(variance_out), Tensor(saved_mean), Tensor(saved_variance), Tensor(reserve_space)
invoke : batch_norm_impl(x, scale, bias, mean, variance, momentum, epsilon, data_layout, is_test, use_global_stats, trainable_statistics, fuse_with_relu)
backward : batch_norm_grad
- api : bce_loss
args : (Tensor input, Tensor label)
output : Tensor
infer_meta :
func : BCELossInferMeta
kernel :
func : bce_loss
backward : bce_loss_grad
# bernoulli
- api : bernoulli
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : bernoulli
# bitwise_and
- api : bitwise_and
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : bitwise_and
# bitwise_not
- api : bitwise_not
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : bitwise_not
# bitwise_or
- api : bitwise_or
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : bitwise_or
# bitwise_xor
- api : bitwise_xor
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : bitwise_xor
# brelu
- api : brelu
args : (Tensor x, float t_min, float t_max)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : brelu
backward : brelu_grad
- api : cast
args : (Tensor x, DataType out_dtype)
output : Tensor
infer_meta :
func : CastInferMeta
kernel :
func : cast
param : [x, out_dtype]
data_type : x
backward : cast_grad
- api : ceil
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : ceil
backward : ceil_grad
- api : celu
args : (Tensor x, float alpha)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : celu
backward : celu_grad
# cholesky
- api : cholesky
args : (Tensor x, bool upper)
output : Tensor
infer_meta :
func : CholeskyInferMeta
kernel :
func : cholesky
backward : cholesky_grad
# cholesky_solve
- api : cholesky_solve
args : (Tensor x, Tensor y, bool upper)
output : Tensor
infer_meta :
func : CholeskySolveInferMeta
kernel :
func : cholesky_solve
backward : cholesky_solve_grad
- api : clip
args : (Tensor x, Scalar(float) min, Scalar(float) max)
output : Tensor(out)
inplace : (x -> out)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : clip
backward : clip_grad
- api : concat
args : (Tensor[] x, Scalar(int64_t) axis)
output : Tensor
infer_meta :
func : ConcatInferMeta
param : [x, axis]
kernel :
func : concat
backward : concat_grad
- api : conj
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : conj
backward : conj_grad
- api : conv2d
args : (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor
invoke : conv2d_impl(input, filter, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search)
backward : conv2d_grad
- api : conv2d_transpose
args : (Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(out)
infer_meta :
func : ConvTransposeInferMeta
kernel :
func : conv2d_transpose
use_gpudnn : true
backward : conv2d_transpose_grad
- api : conv3d
args : (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor
invoke : conv3d_impl(input, filter, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search)
backward : conv3d_grad
- api : conv3d_transpose
args : (Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(out)
infer_meta :
func : ConvTransposeInferMeta
kernel :
func : conv3d_transpose
use_gpudnn : true
backward : conv3d_transpose_grad
- api : copy_to
args : (Tensor x, Place place, bool blocking)
output : Tensor
invoke : copy_to_impl(x, place, blocking)
# cos
- api : cos
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : cos
backward : cos_grad
# cosh
- api : cosh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : cosh
backward : cosh_grad
- api : cross
args : (Tensor x, Tensor y, int axis = 9)
output : Tensor
infer_meta :
func : CrossInferMeta
kernel :
func : cross
backward : cross_grad
# Part of python API paddle.nn.functional.cross_entropy
- api : cross_entropy_with_softmax
args : (Tensor input, Tensor label, bool soft_label, bool use_softmax, bool numeric_stable_mode, int ignore_index, int axis)
output : Tensor(softmax), Tensor(loss)
infer_meta :
func : CrossEntropyWithSoftmaxInferMeta
kernel :
func : cross_entropy_with_softmax
data_type : input
backward : cross_entropy_with_softmax_grad
- api : cumprod
args : (Tensor x, int dim)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : cumprod
backward : cumprod_grad
- api : cumsum
args : (Tensor x, int axis, bool flatten, bool exclusive, bool reverse)
output : Tensor(out)
infer_meta :
func : CumInferMeta
kernel :
func : cumsum
backward : cumsum_grad
- api : deformable_conv
args : (Tensor x, Tensor offset, Tensor filter, Tensor mask, int[] strides, int[] paddings, int[] dilations, int deformable_groups, int groups, int im2col_step)
output : Tensor(out)
infer_meta :
func : DeformableConvInferMeta
kernel :
func : deformable_conv
data_type : x
optional : mask
backward : deformable_conv_grad
- api : depthwise_conv2d
args : (Tensor x, Tensor filter, int[] strides, int[] paddings, str padding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn)
output : Tensor(out)
infer_meta :
func : ConvInferMeta
param : [x, filter, strides, paddings, padding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search]
kernel :
func : depthwise_conv2d
param : [x, filter, strides, paddings, padding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search, fuse_relu]
use_gpudnn : use_gpudnn
backward : depthwise_conv2d_grad
- api : depthwise_conv2d_transpose
args : (Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(out)
infer_meta :
func : ConvTransposeInferMeta
kernel :
func : depthwise_conv2d_transpose
backward : depthwise_conv2d_transpose_grad
- api : det
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : determinant
backward : det_grad
- api : diag
args : (Tensor x, int offset, float padding_value)
output : Tensor
infer_meta :
func : DiagInferMeta
kernel :
func : diag
- api : diagonal
args : (Tensor x, int offset, int axis1, int axis2)
output : Tensor
infer_meta :
func : DiagonalInferMeta
kernel :
func : diagonal
backward : diagonal_grad
- api : digamma
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : digamma
backward : digamma_grad
- api : dist
args : (Tensor x, Tensor y, float p)
output : Tensor
infer_meta :
func : DistInferMeta
kernel :
func : dist
backward : dist_grad
- api : divide
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : divide
backward : divide_grad
- api : dot
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : DotInferMeta
kernel :
func : dot
- api : dropout
args : (Tensor x, Tensor seed_tensor, float p, bool is_test, str mode, int seed, bool fix_seed)
output : Tensor(out), Tensor(mask)
infer_meta :
func : DropoutInferMeta
kernel :
func : dropout
data_type : x
optional : seed_tensor
backward : dropout_grad
# eigh
- api : eigh
args : (Tensor x, str uplo)
output : Tensor(out_w), Tensor(out_v)
infer_meta :
func : EighInferMeta
kernel :
func : eigh
backward : eigh_grad
- api : einsum
args : (Tensor[] x, str equation)
output : Tensor, Tensor[]{x.size()}, Tensor[]{x.size()}
infer_meta :
func : EinsumInferMeta
param : [x, equation]
kernel :
func : einsum
backward : einsum_grad
- api : elementwise_pow
args : (Tensor x, Tensor y)
output : Tensor(out)
infer_meta :
func : ElementwiseInferMeta
kernel :
func : elementwise_pow
backward : elementwise_pow_grad
# elu
- api : elu
args : (Tensor x, float alpha)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : elu
backward : elu_grad
- api : embedding
args : (Tensor x, Tensor weight, int64_t padding_idx=-1, bool sparse=false)
output : Tensor
invoke : embedding_impl(x, weight, padding_idx, sparse)
backward : embedding_grad
- api : empty
args : (IntArray shape, DataType dtype=DataType::FLOAT32, Place place=CPUPlace())
output: Tensor
infer_meta :
func : CreateInferMeta
param : [shape, dtype]
kernel :
func : empty
param : [shape, dtype]
data_type : dtype
backend : place
- api : empty_like
args : (Tensor x, DataType dtype = DataType::UNDEFINED, Place place = {})
output: Tensor
infer_meta :
func : CreateLikeInferMeta
param : [x, dtype]
kernel :
func : empty_like
param : [x, dtype]
data_type : dtype > x
backend : place > x
- api : equal
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : equal
- api : equal_all
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : CompareAllInferMeta
kernel :
func : equal_all
# erf
- api : erf
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : erf
backward : erf_grad
# erfinv
- api : erfinv
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : erfinv
backward : erfinv_grad
# exp
- api : exp
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : exp
backward : exp_grad
# expand
- api : expand
args : (Tensor x, IntArray shape)
output : Tensor
infer_meta :
func : ExpandInferMeta
kernel :
func : expand
backward : expand_grad
# expand_as
- api : expand_as
args : (Tensor x, Tensor y, int[] target_shape)
output : Tensor
infer_meta :
func : ExpandAsInferMeta
kernel :
func : expand_as
optional : y
backward : expand_as_grad
- api : expm1
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : expm1
backward : expm1_grad
- api : eye
args : (int64_t num_rows, int64_t num_columns, DataType dtype=DataType::FLOAT32, Place place={})
output : Tensor(out)
infer_meta :
func : EyeInferMeta
param : [num_rows, num_columns, dtype]
kernel :
func : eye
param : [num_rows, num_columns, dtype]
data_type : dtype
backend : place
- api : flatten
args : (Tensor x, int start_axis, int stop_axis)
output : Tensor(out), Tensor(xshape)
infer_meta :
func : FlattenWithXShapeInferMeta
kernel :
func : flatten_with_xshape
backend : x
inplace : (x -> out)
view : (x -> out)
intermediate : xshape
backward : flatten_grad
# flip
- api : flip
args : (Tensor x, int[] axis)
output : Tensor
infer_meta :
func : FlipInferMeta
kernel :
func : flip
backward : flip_grad
- api : floor
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : floor
backward : floor_grad
- api : floor_divide
args : (Tensor x, Tensor y)
output : Tensor(out)
infer_meta :
func : ElementwiseInferMeta
kernel :
func : floor_divide
- api : fmax
args : (Tensor x, Tensor y, int axis)
output : Tensor(out)
infer_meta :
param: [x, y]
func : ElementwiseInferMeta
kernel :
func : fmax
backward : fmax_grad
- api : fmin
args : (Tensor x, Tensor y, int axis)
output : Tensor(out)
infer_meta :
param: [x, y]
func : ElementwiseInferMeta
kernel :
func : fmin
backward : fmin_grad
- api : frobenius_norm
args : (Tensor x, int64_t[] axis, bool keep_dim, bool reduce_all)
output : Tensor(out)
infer_meta :
func : ReduceInferMetaBase
kernel :
func : frobenius_norm
backward : frobenius_norm_grad
- api : full
args : (IntArray shape, Scalar value, DataType dtype=DataType::FLOAT32, Place place=CPUPlace())
output: Tensor
infer_meta :
func : CreateInferMeta
param : [shape, dtype]
kernel :
func : full
param : [shape, value, dtype]
data_type : dtype
backend : place
- api : full_batch_size_like
args : (Tensor input, int[] shape, DataType dtype, Scalar value, int input_dim_idx, int output_dim_idx, Place place=CPUPlace())
output: Tensor
infer_meta :
func : FullBatchSizeLikeInferMeta
param : [input, shape, value, dtype, input_dim_idx, output_dim_idx]
kernel :
func : full_batch_size_like
param : [input, shape, value, dtype, input_dim_idx, output_dim_idx]
data_type : dtype
backend : place
- api : full_like
args : (Tensor x, Scalar value, DataType dtype = DataType::UNDEFINED, Place place = {})
output: Tensor
infer_meta :
func : CreateLikeInferMeta
param : [x, dtype]
kernel :
func : full_like
param : [x, value, dtype]
data_type : dtype > x
backend : place > x
data_transform :
skip_transform : x
- api : gather
args : (Tensor x, Tensor index, Scalar(int) axis=0)
output : Tensor(out)
infer_meta :
func : GatherInferMeta
kernel :
func : gather
data_type: x
backward : gather_grad
- api : gather_nd
args : (Tensor x, Tensor index)
output : Tensor
infer_meta :
func : GatherNdInferMeta
kernel :
func : gather_nd
data_type : x
backward : gather_nd_grad
- api : gather_tree
args : (Tensor ids, Tensor parents)
output : Tensor
infer_meta :
func : GatherTreeMeta
kernel :
func : gather_tree
- api : gaussian_random
args : (IntArray shape, float mean, float std, int seed, DataType dtype, Place place={})
output: Tensor
infer_meta :
func : GaussianRandomInferMeta
param : [shape, mean, std, seed, dtype]
kernel :
func : gaussian_random
param : [shape, mean, std, seed, dtype]
data_type : dtype
backend : place
- api : gelu
args : (Tensor x, bool approximate)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : gelu
backward : gelu_grad
- api : graph_send_recv
args : (Tensor x, Tensor src_index, Tensor dst_index, str pool_type = "SUM", int64_t out_size = 0)
output : Tensor(out), Tensor(dst_count)
infer_meta :
func : GraphSendRecvInferMeta
kernel :
func : graph_send_recv
data_type : x
intermediate : dst_count
backward : graph_send_recv_grad
- api : greater_equal
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : greater_equal
- api : greater_than
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : greater_than
- api : group_norm
args : (Tensor x, Tensor scale, Tensor bias, float epsilon, int groups, str data_layout)
output : Tensor(y), Tensor(mean), Tensor(variance)
infer_meta :
func : GroupNormInferMeta
kernel :
func : group_norm
optional : scale, bias
intermediate : mean, variance
backward : group_norm_grad
- api : gumbel_softmax
args : (Tensor x, float temperature, bool hard, int axis)
output : Tensor
infer_meta :
func : GumbelSoftmaxInferMeta
kernel :
func : gumbel_softmax
backward : gumbel_softmax_grad
# hard_shrink
- api : hard_shrink
args : (Tensor x, float threshold)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_shrink
backward : hard_shrink_grad
# hard_sigmoid
- api : hard_sigmoid
args : (Tensor x, float slope, float offset)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_sigmoid
backward : hard_sigmoid_grad
- api : hard_swish
args : (Tensor x, float threshold = 6.0, float scale = 6.0, float offset = 3.0)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_swish
backward : hard_swish_grad
# histogram
- api : histogram
args : (Tensor x, int64_t bins, int min, int max)
output : Tensor
infer_meta :
func : HistogramInferMeta
kernel :
func : histogram
- api : huber_loss
args : (Tensor input, Tensor label, float delta)
output : Tensor(out), Tensor(residual)
infer_meta :
func : HuberLossInferMeta
kernel :
func : huber_loss
backward : huber_loss_grad
- api : imag
args : (Tensor x)
output : Tensor
infer_meta :
func : RealAndImagInferMeta
kernel :
func : imag
backward : imag_grad
# increment
- api : increment
args : (Tensor x, float value)
output : Tensor
infer_meta :
func : IncrementInferMeta
kernel :
func : increment
- api : index_sample
args : (Tensor x, Tensor index)
output : Tensor
infer_meta :
func : IndexSampleInferMeta
kernel :
func : index_sample
data_type : x
backward : index_sample_grad
- api : index_select
args : (Tensor x, Tensor index, int dim)
output : Tensor(out)
infer_meta :
func : IndexSelectInferMeta
kernel :
func : index_select
data_type : x
backward : index_select_grad
- api : instance_norm
args : (Tensor x, Tensor scale, Tensor bias, float epsilon)
output : Tensor(y), Tensor(saved_mean), Tensor(saved_variance)
infer_meta :
func : InstanceNormInferMeta
kernel :
func : instance_norm
data_type : x
optional : scale, bias
intermediate : saved_mean, saved_variance
backward : instance_norm_grad
# is_empty
- api : is_empty
args : (Tensor x)
output : Tensor
infer_meta :
func : IsEmptyInferMeta
kernel :
func : is_empty
- api : isclose
args : (Tensor x, Tensor y, Scalar rtol, Scalar atol, bool equal_nan)
output : Tensor(out)
infer_meta :
func : ValueCompareInferMeta
param: [x, y]
kernel :
func : isclose
# isfinite
- api : isfinite
args : (Tensor x)
output : Tensor
infer_meta :
func : IsfiniteInferMeta
kernel :
func : isfinite, infinite_sr
# isinf
- api : isinf
args : (Tensor x)
output : Tensor
infer_meta :
func : IsfiniteInferMeta
kernel :
func : isinf, isinf_sr
# isnan
- api : isnan
args : (Tensor x)
output : Tensor
infer_meta :
func : IsfiniteInferMeta
kernel :
func : isnan, isnan_sr
- api : kldiv_loss
args : (Tensor x, Tensor label, str reduction)
output : Tensor(out)
infer_meta :
func : KLDivInferMeta
kernel :
func : kldiv_loss
data_type : x
backward : kldiv_loss_grad
- api : kron
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : KronInferMeta
kernel :
func : kron
backward : kron_grad
- api : kthvalue
args : (Tensor x, int k, int axis, bool keepdim)
output : Tensor(out), Tensor(indices)
infer_meta :
func : KthvalueInferMeta
kernel :
func : kthvalue
backward : kthvalue_grad
# label_smooth
- api : label_smooth
args : (Tensor label, Tensor prior_dist, float epsilon)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [label]
kernel :
func : label_smooth
data_type : label
optional : prior_dist
backward : label_smooth_grad
- api : layer_norm
args : (Tensor x, Tensor scale, Tensor bias, float epsilon, int begin_norm_axis, bool is_test)
output : Tensor(out), Tensor(mean), Tensor(variance)
infer_meta :
func : LayerNormInferMeta
kernel :
func : layer_norm
data_type : x
backward : layer_norm_grad
optional : scale, bias
# leaky_relu
- api : leaky_relu
args : (Tensor x, float alpha)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : leaky_relu
backward : leaky_relu_grad
- api : lerp
args : (Tensor x, Tensor y, Tensor weight)
output : Tensor
infer_meta :
func : LerpInferMeta
kernel :
func : lerp
backward : lerp_grad
- api : less_equal
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : less_equal
- api : less_than
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : less_than
- api : lgamma
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : lgamma
backward : lgamma_grad
- api : linspace
args : (Tensor start, Tensor stop, Tensor number, DataType dtype)
output : Tensor
infer_meta :
func : LinspaceInferMeta
kernel :
func : linspace
data_type : dtype
- api : log
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : log
backward: log_grad
- api : log10
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : log10
backward: log10_grad
- api : log1p
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : log1p
backward: log1p_grad
- api : log2
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : log2
backward: log2_grad
# log_loss
- api : log_loss
args : (Tensor input, Tensor label, float epsilon)
output : Tensor
infer_meta :
func : LogLossInferMeta
kernel :
func : log_loss
backward : log_loss_grad
- api : log_softmax
args : (Tensor x, int axis)
output : Tensor(out)
infer_meta :
func : UnchangedInferMetaCheckAxis
kernel :
func : log_softmax
backward : log_softmax_grad
- api : logcumsumexp
args : (Tensor x, int axis, bool flatten, bool exclusive, bool reverse)
output : Tensor(out)
infer_meta :
func : CumInferMeta
kernel :
func : logcumsumexp
backward : logcumsumexp_grad
# logical_and
- api : logical_and
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : logical_and
# logical_not
- api : logical_not
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : logical_not
# logical_or
- api : logical_or
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : logical_or
# logical_xor
- api : logical_xor
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : logical_xor
# logit
- api : logit
args : (Tensor x, float eps = 1e-6f)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : logit
backward : logit_grad
# logsigmoid
- api : logsigmoid
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : logsigmoid
backward : logsigmoid_grad
- api : logsumexp
args : (Tensor x, int64_t[] axis, bool keepdim, bool reduce_all)
output : Tensor(out)
infer_meta :
func : LogsumexpInferMeta
kernel :
func : logsumexp
backward : logsumexp_grad
# masked_select
- api : masked_select
args : (Tensor x, Tensor mask)
output : Tensor
infer_meta :
func : MaskedSelectInferMeta
kernel :
func : masked_select
data_type : x
backward : masked_select_grad
- api : matmul
args : (Tensor x, Tensor y, bool transpose_x = false, bool transpose_y = false)
output : Tensor
infer_meta :
func : MatmulInferMeta
kernel :
func : matmul
backward : matmul_grad
# matrix_power
- api : matrix_power
args : (Tensor x, int n)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : matrix_power
backward : matrix_power_grad
- api : matrix_rank
args : (Tensor x, float tol, bool use_default_tol=true, bool hermitian=false)
output : Tensor(out)
infer_meta :
func : MatrixRankInferMeta
param : [x, use_default_tol, hermitian]
kernel :
func : matrix_rank
- api : matrix_rank_tol
args : (Tensor x, Tensor atol_tensor, bool use_default_tol=true, bool hermitian=false)
output : Tensor(out)
infer_meta :
func : MatrixRankTolInferMeta
kernel :
func : matrix_rank_tol
- api : max
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : max
backward : max_grad
- api : max_pool2d_with_index
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
output : Tensor(out), Tensor(mask)
infer_meta :
func : MaxPoolWithIndexInferMeta
kernel :
func : max_pool2d_with_index
backward : max_pool2d_with_index_grad
- api : max_pool3d_with_index
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
output : Tensor(out), Tensor(mask)
infer_meta :
func : MaxPoolWithIndexInferMeta
kernel :
func : max_pool3d_with_index
backward : max_pool3d_with_index_grad
- api : maximum
args : (Tensor x, Tensor y)
output : Tensor(out)
infer_meta :
func : ElementwiseInferMeta
kernel :
func : maximum
backward : maximum_grad
- api : maxout
args : (Tensor x, int groups, int axis)
output : Tensor(out)
infer_meta :
func : MaxOutInferMeta
kernel :
func : maxout
backward : maxout_grad
- api : mean
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : mean
backward : mean_grad
- api : mean_all
args : (Tensor x)
output : Tensor
infer_meta :
func : MeanAllInferMeta
kernel :
func : mean_all
backward : mean_all_grad
- api : meshgrid
args : (Tensor[] inputs)
output : Tensor[]{inputs.size()}
infer_meta :
func : MeshgridInferMeta
kernel :
func : meshgrid
backward : meshgrid_grad
- api : min
args : (Tensor x, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : ReduceInferMeta
kernel :
func : min
backward : min_grad
- api : minimum
args : (Tensor x, Tensor y)
output : Tensor(out)
infer_meta :
func : ElementwiseInferMeta
kernel :
func : minimum
backward : minimum_grad
- api : mish
args : (Tensor x, float lambda)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : mish
backward : mish_grad
- api : mode
args : (Tensor x, int axis, bool keepdim)
output : Tensor(out), Tensor(indices)
infer_meta :
func : ModeInferMeta
kernel :
func : mode
backward : mode_grad
- api : modulo
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : modulo
backward : modulo_grad
- api : momentum
args : (Tensor param, Tensor grad, Tensor velocity, Tensor learning_rate, Tensor master_param, float mu, bool use_nesterov = false, str regularization_method = "", float regularization_coeff = 0.0, bool multi_precision = false, float rescale_grad = 1.0f)
output : Tensor(param_out), Tensor(velocity_out), Tensor(master_param_out)
invoke : momentum_impl(param, grad, velocity, learning_rate, master_param, mu, use_nesterov, regularization_method, regularization_coeff, multi_precision, rescale_grad)
optional : master_param
- api : multi_dot
args : (Tensor[] x)
output : Tensor
infer_meta :
func : MultiDotInferMeta
kernel :
func : multi_dot
backward : multi_dot_grad
# multinomial
- api : multinomial
args : (Tensor x, int num_samples, bool replacement)
output : Tensor
infer_meta :
func : MultinomialInferMeta
kernel :
func : multinomial
- api : multiplex
args : (Tensor[] ins, Tensor ids)
output : Tensor
infer_meta :
func : MultiplexInferMeta
kernel :
func : multiplex
data_type : ins
backward : multiplex_grad
- api : multiply
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : multiply
backward : multiply_grad
- api : mv
args : (Tensor x, Tensor vec)
output : Tensor
infer_meta :
func : MvInferMeta
kernel :
func : mv
backward : mv_grad
- api : nll_loss
args : (Tensor input, Tensor label, Tensor weight, int64_t ignore_index, str reduction)
output : Tensor(out), Tensor(total_weight)
infer_meta :
func : NllLossRawInferMeta
kernel :
func : nll_loss
data_type : input
optional : weight
backward : nll_loss_grad
- api : norm
args : (Tensor x, int axis, float epsilon, bool is_test)
output : Tensor(out), Tensor(norm)
infer_meta :
func : NormInferMeta
kernel :
func : norm
intermediate : norm
backward : norm_grad
- api : not_equal
args : (Tensor x, Tensor y, int axis = -1)
output : Tensor
infer_meta :
func : CompareInferMeta
kernel :
func : not_equal
- api : one_hot
args : (Tensor x, Scalar(int) num_classes)
output : Tensor
infer_meta :
func : OneHotInferMeta
kernel :
func : one_hot
- api : ones_like
args : (Tensor x, DataType dtype=DataType::UNDEFINED, Place place={})
output : Tensor
invoke : full_like(x, 1, dtype, place)
- api : p_norm
args : (Tensor x, float porder, int axis, float epsilon, bool keepdim, bool asvector=false)
output : Tensor(out)
infer_meta :
func : PNormInferMeta
kernel :
func : p_norm
backward : p_norm_grad
# pad
- api : pad
args : (Tensor x, int[] paddings, float pad_value)
output : Tensor
infer_meta :
func : PadInferMeta
kernel :
func : pad
backward : pad_grad
- api : pad3d
args : (Tensor x, IntArray paddings, str mode, float pad_value, str data_format)
output : Tensor(out)
infer_meta :
func : Pad3dInferMeta
kernel :
func : pad3d
backward : pad3d_grad
# pixel_shuffle
- api : pixel_shuffle
args : (Tensor x, int upscale_factor, str data_format)
output : Tensor
infer_meta :
func : PixelShuffleInferMeta
kernel :
func : pixel_shuffle
backward : pixel_shuffle_grad
# poisson
- api : poisson
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : poisson
backward : poisson_grad
- api : pool2d
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(out)
infer_meta :
func : PoolInferMeta
kernel :
func : pool2d
use_gpudnn : true
backward : pool2d_grad
# Used in adaptive_avg_pool2d API
- api : pool2d_gpudnn_unused
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(out)
infer_meta :
func : PoolInferMeta
kernel :
func : pool2d
use_gpudnn : false
backward : pool2d_grad_gpudnn_unused
- api : pool3d
args : (Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(out)
infer_meta :
func : PoolInferMeta
kernel :
func : pool3d
use_gpudnn : true
backward : pool3d_grad
- api : pow
args : (Tensor x, Scalar s)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : pow
backward : pow_grad
- api : prelu
args : (Tensor x, Tensor alpha, str data_format, str mode)
output : Tensor(out)
infer_meta :
func : PReluInferMeta
kernel :
func : prelu
backward : prelu_grad
- api : psroi_pool
args : (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, int output_channels, float spatial_scale)
output : Tensor
infer_meta :
func : PsroiPoolInferMeta
kernel :
func : psroi_pool
data_type : x
optional : boxes_num
backward : psroi_pool_grad
# put_along_axis
- api : put_along_axis
args : (Tensor x, Tensor index, Tensor value, int axis, str reduce)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [index]
kernel :
func : put_along_axis
data_type : x
backward : put_along_axis_grad
- api : qr
args : (Tensor x, str mode)
output : Tensor(q), Tensor(r)
infer_meta :
func : QrInferMeta
kernel :
func : qr
# backward : qr_grad
- api : randint
args : (int low, int high, IntArray shape, DataType dtype=DataType::INT64, Place place={})
output : Tensor(out)
infer_meta :
func : RandintInferMeta
param : [low, high, shape, dtype]
kernel :
func : randint
param : [low, high, shape, dtype]
data_type : dtype
backend : place
- api : randperm
args : (int n, DataType dtype, Place place={})
output : Tensor
infer_meta :
func : RandpermInferMeta
param : [n, dtype]
kernel :
func : randperm
param : [n, dtype]
data_type : dtype
backend : place
- api : real
args : (Tensor x)
output : Tensor
infer_meta :
func : RealAndImagInferMeta
kernel :
func : real
backward : real_grad
- api : reciprocal
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : reciprocal
backward : reciprocal_grad
# reduce_prod
- api : reduce_prod
args : (Tensor x, int64_t[] dims, bool keep_dim, bool reduce_all)
output : Tensor
infer_meta :
func : ReduceInferMetaBase
kernel :
func : prod_raw
backward : reduce_prod_grad
- api : relu
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : relu
inplace : (x -> out)
backward : relu_grad
- api : reshape
args : (Tensor x, IntArray shape)
output : Tensor(out), Tensor(xshape)
infer_meta :
func : ReshapeWithXShapeInferMeta
kernel :
func : reshape_with_xshape
inplace : (x -> out)
view: (x -> out)
intermediate : xshape
backward: reshape_grad
- api : roi_align
args : (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale, int sampling_ratio, bool aligned)
output : Tensor
infer_meta :
func : RoiAlignInferMeta
kernel :
func : roi_align
data_type : x
optional : boxes_num
backward : roi_align_grad
- api : roi_pool
args : (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale)
output : Tensor(out), Tensor(arg_max)
infer_meta :
func : RoiPoolInferMeta
kernel :
func : roi_pool
data_type : x
optional : boxes_num
intermediate : arg_max
backward : roi_pool_grad
- api : roll
args : (Tensor x, IntArray shifts, int64_t[] axis)
output : Tensor(out)
infer_meta :
func : RollInferMeta
kernel :
func : roll
backward : roll_grad
- api : round
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : round
backward : round_grad
- api : rsqrt
args : (Tensor x)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
kernel :
func : rsqrt
inplace : (x -> out)
backward : rsqrt_grad
- api : scale
args : (Tensor x, Scalar scale, float bias, bool bias_after_scale)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : scale, scale_sr
inplace : (x -> out)
backward : scale_grad
- api : scatter
args : (Tensor x, Tensor index, Tensor updates, bool overwrite)
output : Tensor
infer_meta :
func : ScatterInferMeta
dtype : x
kernel :
func : scatter
backward : scatter_grad
- api : scatter_nd_add
args : (Tensor x, Tensor index, Tensor updates)
output : Tensor
infer_meta :
func : ScatterNdAddInferMeta
dtype : x
kernel :
func : scatter_nd_add
backward : scatter_nd_add_grad
- api : searchsorted
args : (Tensor sorted_sequence, Tensor value, bool out_int32, bool right)
output : Tensor(out)
infer_meta :
func : SearchsortedInferMeta
kernel :
func : searchsorted
data_type : sorted_sequence
# segment_pool
- api : segment_pool
args : (Tensor x, Tensor segment_ids, str pooltype)
output : Tensor(out), Tensor(summed_ids)
infer_meta :
func : SegmentPoolInferMeta
kernel :
func : segment_pool
data_type : x
backward : segment_pool_grad
# selu
- api : selu
args : (Tensor x, float scale, float alpha)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : selu
backward : selu_grad
- api : sgd
args : (Tensor param, Tensor learning_rate, Tensor grad, Tensor master_param, bool multi_precision)
output : Tensor(param_out), Tensor(master_param_out)
invoke : sgd_impl(param, learning_rate, grad, master_param, multi_precision)
optional : master_param
- api : shape
args : (Tensor input)
output : Tensor
infer_meta :
func : ShapeInferMeta
kernel :
func : shape, shape_sr
data_transform:
skip_transform : input
# shard_index
- api : shard_index
args : (Tensor in, int index_num, int nshards, int shard_id, int ignore_value)
output : Tensor
infer_meta :
func : ShardIndexInferMeta
kernel :
func : shard_index
# sigmoid
- api : sigmoid
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sigmoid
backward : sigmoid_grad
# sigmoid_cross_entropy_with_logits
- api : sigmoid_cross_entropy_with_logits
args : (Tensor x, Tensor label, bool normalize, int ignore_index)
output : Tensor
infer_meta :
func : SigmoidCrossEntropyWithLogitsInferMeta
kernel :
func : sigmoid_cross_entropy_with_logits
backward : sigmoid_cross_entropy_with_logits_grad
- api : sign
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sign
# silu
- api : silu
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : silu
backward : silu_grad
# sin
- api : sin
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sin
backward : sin_grad
# sinh
- api : sinh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sinh
backward : sinh_grad
# size
- api : size
args : (Tensor x)
output : Tensor
infer_meta :
func : SizeInferMeta
kernel :
func : size
data_transform:
skip_transform : x
- api : slice
args : (Tensor input, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis)
output : Tensor
infer_meta :
func : SliceRawInferMeta
kernel :
func : slice
backward : slice_grad
# soft_shrink
- api : soft_shrink
args : (Tensor x, float lambda)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : soft_shrink
backward : soft_shrink_grad
- api : softmax
args : (Tensor x, int axis)
output : Tensor
infer_meta :
func : SoftmaxInferMeta
kernel :
func : softmax
use_gpudnn : true
backward : softmax_grad
- api : split
args : (Tensor x, IntArray num_or_sections, Scalar(int) axis)
output : Tensor[]
invoke : split_impl(x, num_or_sections, axis)
backward : split_grad
- api : sqrt
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : sqrt
backward : sqrt_grad
- api : square
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : square
backward : square_grad
- api : squeeze
args : (Tensor x, int[] axes)
output : Tensor(out), Tensor(xshape)
infer_meta :
func : SqueezeInferMeta
kernel :
func : squeeze
view: (x -> out)
intermediate : xshape
backward : squeeze_grad
- api : stack
args : (Tensor[] x, int axis)
output : Tensor
infer_meta :
func : StackInferMeta
kernel :
func : stack
backward : stack_grad
- api : strided_slice
args : (Tensor x, int[] axes, IntArray starts, IntArray ends, IntArray strides)
output : Tensor
infer_meta :
func : StridedSliceInferMeta
kernel :
func : strided_slice
backward : strided_slice_grad
- api : subtract
args : (Tensor x, Tensor y)
output : Tensor
infer_meta :
func : ElementwiseInferMeta
kernel :
func : subtract
backward : subtract_grad
- api : sum
args : (Tensor x, int64_t[] dims={}, DataType out_dtype=DataType::UNDEFINED, bool keep_dim=false)
output : Tensor(out)
infer_meta :
func : SumInferMeta
kernel :
func : sum
data_type : x
backward : sum_grad
# The python API paddle.nn.functional.swish has no `bete` argument, it may be removed later
- api : swish
args : (Tensor x, float beta=1.0)
output : Tensor(out)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : swish
backward : swish_grad
# take_along_axis
- api : take_along_axis
args : (Tensor x, Tensor index, int axis)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [index]
kernel :
func : take_along_axis
data_type : x
backward : take_along_axis_grad
# tan
- api : tan
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : tan
backward : tan_grad
# tanh
- api : tanh
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : tanh
backward : tanh_grad
# tanh_shrink
- api : tanh_shrink
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : tanh_shrink
backward : tanh_shrink_grad
# thresholded_relu
- api : thresholded_relu
args : (Tensor x, float threshold)
output : Tensor
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : thresholded_relu
backward : thresholded_relu_grad
# tile
- api : tile
args : (Tensor x, IntArray repeat_times)
output : Tensor
infer_meta :
func : TileInferMeta
kernel :
func : tile
backward : tile_grad
- api : top_k
args : (Tensor x, Scalar k, int axis = -1, bool largest = true, bool sorted = true)
output : Tensor(out), Tensor(indices)
infer_meta :
func : TopKInferMeta
kernel :
func : top_k
backward : top_k_grad
- api : trace
args : (Tensor x, int offset, int axis1, int axis2)
output : Tensor
infer_meta :
func : TraceInferMeta
kernel :
func : trace
backward : trace_grad
- api : transpose
args : (Tensor x, int[] axis)
output : Tensor
infer_meta :
func : TransposeInferMeta
kernel :
func : transpose
backward : transpose_grad
- api : triangular_solve
args : (Tensor x, Tensor y, bool upper, bool transpose, bool unitriangular)
output : Tensor
infer_meta :
func : TriangularSolveInferMeta
kernel :
func : triangular_solve
backward : triangular_solve_grad
- api : tril_indices
args : (int rows, int cols, int offset, DataType dtype, Place place={})
output : Tensor(out)
infer_meta :
func : TrilIndicesInferMeta
param : [rows, cols, offset, dtype]
kernel :
func : tril_indices
param : [rows, cols, offset, dtype]
data_type : dtype
backend : place
- api : tril_triu
args : (Tensor x, int diagonal, bool lower)
output : Tensor(out)
infer_meta :
func : TrilTriuInferMeta
kernel :
func : tril_triu
backward : tril_triu_grad
- api : trunc
args : (Tensor x)
output : Tensor
infer_meta :
func : UnchangedInferMeta
kernel :
func : trunc
backward : trunc_grad
# python API: paddle.nn.initializer.TruncatedNormal
- api : truncated_gaussian_random
args : (int[] shape, float mean, float std, int seed, DataType dtype=DataType::FLOAT32, Place place={})
output : Tensor
infer_meta :
func : TruncatedGaussianRandomInferMeta
param : [shape, mean, std, seed, dtype]
kernel :
func : truncated_gaussian_random
param : [shape, mean, std, seed, dtype]
backend : place
data_type : dtype
- api : unbind
args : (Tensor input, int axis)
output : Tensor[] {axis<0 ? input.dims()[input.dims().size()+axis]:input.dims()[axis]}
infer_meta :
func : UnbindInferMeta
kernel :
func : unbind
backward : unbind_grad
# unfold
- api : unfold
args : (Tensor x, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations)
output : Tensor
infer_meta :
func : UnfoldInferMeta
kernel :
func : unfold
backward : unfold_grad
- api : uniform_random
args : (IntArray shape, DataType dtype, float min, float max, int seed, Place place={})
output : Tensor(out)
infer_meta :
func : UniformRandomInferMeta
param: [shape, dtype, min, max, seed]
kernel :
func : uniform_random
param: [shape, dtype, min, max, seed]
data_type : dtype
backend : place
# The `axis` argument of Python API paddle.unique is not vector
- api : unique
args : (Tensor x, bool return_index, bool return_inverse, bool return_counts, int[] axis, DataType dtype=DataType::INT64)
output : Tensor(out), Tensor(indices), Tensor(inverse), Tensor(counts)
infer_meta :
func : UniqueInferMeta
kernel :
func : unique
data_type : x
- api : unsqueeze
args : (Tensor x, IntArray axis)
output : Tensor(out), Tensor(xshape)
infer_meta :
func : UnsqueezeInferMeta
kernel :
func : unsqueeze
view: (x -> out)
intermediate : xshape
backward : unsqueeze_grad
# viterbi_decode
- api : viterbi_decode
args : (Tensor input, Tensor transition, Tensor length, bool include_bos_eos_tag)
output : Tensor(scores), Tensor(path)
infer_meta :
func : ViterbiDecodeInferMeta
kernel :
func : viterbi_decode
data_type : input
- api : where
args : (Tensor condition, Tensor x, Tensor y)
output : Tensor
infer_meta :
func : WhereInferMeta
kernel :
func : where
backward : where_grad
# where_index
- api : where_index
args : (Tensor condition)
output : Tensor
infer_meta :
func : WhereIndexInferMeta
kernel :
func : where_index
# yolo_box
- api : yolo_box
args : (Tensor x, Tensor img_size, int[] anchors, int class_num, float conf_thresh, int downsample_ratio, bool clip_bbox, float scale_x_y=1.0, bool iou_aware=false, float iou_aware_factor=0.5)
output : Tensor(boxes), Tensor(scores)
infer_meta :
func : YoloBoxInferMeta
kernel :
func : yolo_box
data_type : x
- api : zeros_like
args : (Tensor x, DataType dtype=DataType::UNDEFINED, Place place = {})
output : Tensor
invoke : full_like(x, 0, dtype, place)
python/paddle/utils/code_gen/legacy_backward.yaml 0 → 100644
浏览文件 @ fcd32950
- backward_api : abs_double_grad
forward : abs_grad (Tensor x, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_x_grad)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : abs_double_grad
data_transform:
skip_transform : grad_x_grad
- backward_api : abs_grad
forward : abs (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : abs_grad
data_transform:
skip_transform : out_grad
backward : abs_double_grad
- backward_api : acos_grad
forward : acos (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : acos_grad
inplace : (out_grad -> x_grad)
- backward_api : acosh_grad
forward : acosh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : acosh_grad
inplace : (out_grad -> x_grad)
- backward_api : add_double_grad
forward : add_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [grad_out]
kernel :
func : add_double_grad
optional : grad_x_grad, grad_y_grad
backward : add_triple_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : add_grad
forward : add (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : add_grad
no_need_buffer : x, y
backward : add_double_grad
inplace : (out_grad -> x_grad)
- backward_api : add_n_grad
forward : add_n (Tensor[] x) -> Tensor(out)
args : (Tensor[] x, Tensor out_grad)
output : Tensor[](x_grad){x.size()}
invoke : add_n_grad_impl(x, out_grad, x_grad)
no_need_buffer : x
- backward_api : add_triple_grad
forward : add_double_grad (Tensor y, Tensor grad_out, Tensor grad_grad_x, Tensor grad_grad_y, int axis = -1) -> Tensor(grad_grad_out)
args : (Tensor grad_grad_x, Tensor grad_grad_y, Tensor grad_grad_out_grad, int axis = -1)
output : Tensor(grad_grad_x_grad), Tensor(grad_grad_y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [grad_grad_x, grad_grad_y]
kernel :
func : add_triple_grad
inplace : (grad_grad_out_grad -> grad_grad_x_grad)
- backward_api : addmm_grad
forward : addmm (Tensor input, Tensor x, Tensor y, float alpha, float beta) -> Tensor(out)
args : (Tensor input, Tensor x, Tensor y, Tensor out_grad, float alpha, float beta)
output : Tensor(input_grad), Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [input, x, y]
kernel :
func : addmm_grad
- backward_api : argsort_grad
forward : argsort (Tensor x, int axis, bool descending) -> Tensor(out), Tensor(indices)
args : (Tensor indices, Tensor x, Tensor out_grad, int axis, bool descending)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : argsort_grad
no_need_buffer : x
- backward_api : asin_grad
forward : asin (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : asin_grad
inplace : (out_grad -> x_grad)
- backward_api : asinh_grad
forward : asinh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : asinh_grad
inplace : (out_grad -> x_grad)
- backward_api : assign_grad
forward : assign (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
kernel :
func : assign
inplace : (out_grad -> x_grad)
- backward_api : assign_out__grad
forward : assign_out_ (Tensor x, Tensor output) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
kernel :
func : assign
inplace : (out_grad -> x_grad)
- backward_api : atan2_grad
forward : atan2 (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : atan2_grad
- backward_api : atan_grad
forward : atan (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : atan_grad
inplace : (out_grad -> x_grad)
- backward_api : atanh_grad
forward : atanh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : atanh_grad
inplace : (out_grad -> x_grad)
- backward_api : batch_norm_double_grad
forward : batch_norm_grad (Tensor x, Tensor scale, Tensor bias, Tensor out_mean, Tensor out_variance, Tensor saved_mean, Tensor saved_variance, Tensor reserve_space, Tensor grad_out, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu) -> Tensor(grad_x), Tensor(grad_scale), Tensor(grad_bias)
args : (Tensor x, Tensor scale, Tensor out_mean, Tensor out_variance, Tensor saved_mean, Tensor saved_variance, Tensor grad_out, Tensor grad_x_grad, Tensor grad_scale_grad, Tensor grad_bias_grad, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [x, scale, x]
kernel :
func : batch_norm_grad_grad
data_type : x
optional : out_mean, out_variance
inplace : (grad_out -> grad_out_grad)
- backward_api : batch_norm_grad
forward : batch_norm (Tensor x, Tensor scale, Tensor bias, Tensor mean, Tensor variance, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu) -> Tensor(out), Tensor(mean_out), Tensor(variance_out), Tensor(saved_mean), Tensor(saved_variance), Tensor(reserve_space)
args : (Tensor x, Tensor scale, Tensor bias, Tensor mean_out, Tensor variance_out, Tensor saved_mean, Tensor saved_variance, Tensor reserve_space, Tensor out_grad, float momentum, float epsilon, str data_layout, bool is_test, bool use_global_stats, bool trainable_statistics, bool fuse_with_relu)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [x, scale, bias]
kernel :
func : batch_norm_grad
data_type : out_grad
optional : mean_out, variance_out, reserve_space
backward : batch_norm_double_grad
- backward_api : bce_loss_grad
forward : bce_loss (Tensor input, Tensor label) -> Tensor(out)
args : (Tensor input, Tensor label, Tensor out_grad)
output : Tensor(input_grad)
infer_meta :
func : UnchangedInferMeta
param : [input]
kernel :
func : bce_loss_grad
inplace : (out_grad -> input_grad)
- backward_api : brelu_grad
forward : brelu (Tensor x, float t_min, float t_max) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float t_min, float t_max)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : brelu_grad
inplace : (out_grad -> x_grad)
- backward_api : cast_grad
forward : cast (Tensor x, DataType out_dtype) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : cast_grad
data_type : out_grad
no_need_buffer : x
- backward_api : ceil_grad
forward : ceil(Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : ceil_grad
inplace : (out_grad -> x_grad)
- backward_api : celu_double_grad
forward : celu_grad(Tensor x, Tensor grad_out, float alpha) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_out, Tensor grad_x_grad, float alpha)
output : Tensor(x_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, x]
kernel :
func : celu_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : celu_grad
forward : celu(Tensor x, float alpha) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float alpha)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : celu_grad
backward : celu_double_grad
inplace : (out_grad -> x_grad)
- backward_api : cholesky_grad
forward : cholesky (Tensor x, bool upper) -> Tensor(out)
args : (Tensor out, Tensor out_grad, bool upper)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : cholesky_grad
- backward_api : cholesky_solve_grad
forward : cholesky_solve (Tensor x, Tensor y, bool upper) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, bool upper)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : cholesky_solve_grad
- backward_api : clip_double_grad
forward : clip_grad (Tensor x, Tensor grad_out, Scalar min = 0., Scalar max = 0.) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_x_grad, Scalar min = 0., Scalar max = 0.)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : clip_grad
- backward_api : clip_grad
forward : clip (Tensor x, Scalar min, Scalar max) -> Tensor(out)
args : (Tensor x, Tensor out_grad, Scalar min = 0., Scalar max = 0.)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : clip_grad
backward : clip_double_grad
inplace : (out_grad -> x_grad)
- backward_api : concat_double_grad
forward : concat_grad (Tensor[] x, Tensor grad_out, Scalar axis) -> Tensor[](grad_x)
args : (Tensor[] grad_x_grad, Scalar axis = 0)
output : Tensor(grad_out_grad)
infer_meta :
func : ConcatInferMeta
param : [grad_x_grad, axis]
kernel :
func : concat
- backward_api : concat_grad
forward : concat (Tensor[] x, Scalar axis) -> Tensor(out)
args : (Tensor[] x, Tensor out_grad, Scalar axis = 0)
output : Tensor[](x_grad){x.size()}
infer_meta :
func : UnchangedMultiInferMeta
param : [x]
kernel :
func : concat_grad
no_need_buffer : x
backward : concat_double_grad
- backward_api : conj_grad
forward : conj (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : conj
- backward_api : conv2d_grad
forward : conv2d (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(out)
args : (Tensor input, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor(input_grad), Tensor(filter_grad)
invoke : conv2d_grad_impl(input, filter, out_grad, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search, input_grad, filter_grad)
backward : conv2d_grad_grad
- backward_api : conv2d_grad_grad
forward : conv2d_grad (Tensor input, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(grad_input), Tensor(grad_filter)
args : (Tensor input, Tensor filter, Tensor grad_out, Tensor grad_input_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor(input_grad), Tensor(filter_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param: [input, filter, grad_out]
kernel :
func : conv2d_grad_grad
use_gpudnn : true
optional : grad_input_grad, grad_filter_grad
- backward_api : conv2d_transpose_double_grad
forward : conv2d_transpose_grad(Tensor x, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(grad_x), Tensor(grad_filter)
args : (Tensor x, Tensor filter, Tensor grad_out, Tensor grad_x_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(x_grad), Tensor(filter_grad), Tensor(grad_out_grad)
infer_meta :
func : Conv2dTransposeDoubleGradInferMeta
kernel :
func : conv2d_transpose_grad_grad
use_gpudnn : true
- backward_api : conv2d_transpose_grad
forward : conv2d_transpose(Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(out)
args : (Tensor x, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(x_grad), Tensor(filter_grad)
infer_meta :
func : ConvTransposeGradInferMeta
kernel :
func : conv2d_transpose_grad
use_gpudnn : true
backward : conv2d_transpose_double_grad
- backward_api : conv3d_grad
forward : conv3d (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(out)
args : (Tensor input, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor(input_grad), Tensor(filter_grad)
invoke : conv3d_grad_impl(input, filter, out_grad, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search, input_grad, filter_grad)
backward : conv3d_grad_grad
- backward_api : conv3d_grad_grad
forward : conv3d_grad (Tensor input, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search) -> Tensor(grad_input), Tensor(grad_filter)
args : (Tensor input, Tensor filter, Tensor grad_out, Tensor grad_input_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search)
output : Tensor(input_grad), Tensor(filter_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param: [input, filter, grad_out]
kernel :
func : conv3d_grad_grad
use_gpudnn : true
optional : grad_input_grad, grad_filter_grad
- backward_api : conv3d_transpose_grad
forward : conv3d_transpose(Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(out)
args : (Tensor x, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(x_grad), Tensor(filter_grad)
infer_meta :
func : ConvTransposeGradInferMeta
kernel :
func : conv3d_transpose_grad
use_gpudnn : true
- backward_api : cos_grad
forward : cos (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : cos_grad
inplace : (out_grad -> x_grad)
- backward_api : cosh_grad
forward : cosh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : cosh_grad
inplace : (out_grad -> x_grad)
- backward_api : cross_entropy_with_softmax_grad
forward : cross_entropy_with_softmax (Tensor input, Tensor label, bool soft_label, bool use_softmax, bool numeric_stable_mode, int ignore_index, int axis) -> Tensor(softmax), Tensor(loss)
args : (Tensor label, Tensor softmax, Tensor loss_grad, bool soft_label, bool use_softmax, bool numeric_stable_mode, int ignore_index, int axis)
output : Tensor(input_grad)
infer_meta :
func : CrossEntropyWithSoftmaxGradInferMeta
kernel :
func : cross_entropy_with_softmax_grad
data_type : softmax
inplace : (softmax -> input_grad)
- backward_api : cross_grad
forward : cross (Tensor x, Tensor y, int axis = 9) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : cross_grad
- backward_api : cumprod_grad
forward : cumprod (Tensor x, int dim) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int dim)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : cumprod_grad
- backward_api : cumsum_grad
forward : cumsum(Tensor x, int axis, bool flatten, bool exclusive, bool reverse) -> Tensor(out)
infer_meta :
func : UnchangedInferMeta
param : [x]
args : (Tensor out_grad, int axis, bool flatten, bool exclusive, bool reverse)
output : Tensor(x_grad)
invoke : cumsum(out_grad, axis, flatten, exclusive, !reverse)
- backward_api : deformable_conv_grad
forward : deformable_conv(Tensor x, Tensor offset, Tensor filter, Tensor mask, int[] strides, int[] paddings, int[] dilations, int deformable_groups, int groups, int im2col_step) -> Tensor(out)
args : (Tensor x, Tensor offset, Tensor filter, Tensor mask, Tensor out_grad, int[] strides, int[] paddings, int[] dilations, int deformable_groups, int groups, int im2col_step)
output : Tensor(x_grad), Tensor(offset_grad), Tensor(filter_grad), Tensor(mask_grad)
infer_meta :
func : DeformableConvGradInferMeta
kernel :
func : deformable_conv_grad
data_type : x
optional : mask
- backward_api : depthwise_conv2d_grad
forward : depthwise_conv2d (Tensor input, Tensor filter, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn) -> Tensor(out)
args : (Tensor input, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn)
output : Tensor(input_grad), Tensor(filter_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [input, filter]
kernel :
func : depthwise_conv2d_grad
param : [input, filter, out_grad, strides, paddings, paddding_algorithm, groups, dilations, data_format, use_addto, workspace_size_MB, exhaustive_search, fuse_relu]
use_gpudnn : use_gpudnn
backward : depthwise_conv2d_grad_grad
- backward_api : depthwise_conv2d_grad_grad
forward : depthwise_conv2d_grad (Tensor input, Tensor filter, Tensor grad_out, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu, bool use_gpudnn) -> Tensor(grad_input), Tensor(grad_filter)
args : (Tensor input, Tensor filter, Tensor grad_out, Tensor grad_input_grad, Tensor grad_filter_grad, int[] strides, int[] paddings, str paddding_algorithm, int groups, int[] dilations, str data_format, bool use_addto, int workspace_size_MB, bool exhaustive_search, bool fuse_relu)
output : Tensor(input_grad), Tensor(filter_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param: [input, filter, grad_out]
kernel :
func : depthwise_conv2d_grad_grad
optional : grad_input_grad, grad_filter_grad
- backward_api : depthwise_conv2d_transpose_grad
forward : depthwise_conv2d_transpose(Tensor x, Tensor filter, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format) -> Tensor(out)
args : (Tensor x, Tensor filter, Tensor out_grad, int[] strides, int[] paddings, int[] output_padding, int[] output_size, str padding_algorithm, int groups, int[] dilations, str data_format)
output : Tensor(x_grad), Tensor(filter_grad)
infer_meta :
func : ConvTransposeGradInferMeta
kernel :
func : depthwise_conv2d_transpose_grad
- backward_api : det_grad
forward : det (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : determinant_grad
- backward_api : diagonal_grad
forward : diagonal (Tensor x, int offset, int axis1, int axis2) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int offset = 0, int axis1 = 0, int axis2 = 1)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : diagonal_grad
no_need_buffer : x
- backward_api : digamma_grad
forward : digamma (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : digamma_grad
- backward_api : dist_grad
forward : dist (Tensor x, Tensor y, float p) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, float p)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : dist_grad
- backward_api : divide_double_grad
forward : divide_grad (Tensor x, Tensor y, Tensor out, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor y, Tensor out, Tensor grad_x, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
output : Tensor(y_grad), Tensor(out_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [y, grad_x, grad_x]
kernel :
func : divide_double_grad
data_type : out
optional : grad_x_grad, grad_y_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : divide_grad
forward : divide (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : divide_grad
backward : divide_double_grad
- backward_api : dropout_grad
forward : dropout (Tensor x, Tensor seed_tensor, float p, bool is_test, str mode, int seed, bool fix_seed) -> Tensor(out), Tensor(mask)
args : (Tensor mask, Tensor out_grad, float p, bool is_test, str mode)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : dropout_grad
- backward_api : eigh_grad
forward : eigh (Tensor x, str uplo) -> Tensor(out_w), Tensor(out_v)
args : (Tensor out_w, Tensor out_v, Tensor out_w_grad, Tensor out_v_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_v]
kernel :
func : eigh_grad
data_type : out_v
data_transform:
skip_transform : out_w, out_w_grad
- backward_api : einsum_grad
forward : einsum (Tensor[] x, str equation) -> Tensor(out), Tensor[](inner_cache), Tensor[](x_shape)
args : (Tensor[] x_shape, Tensor[] inner_cache, Tensor out_grad, str equation)
output : Tensor[](x_grad){x.size()}
infer_meta :
func : UnchangedMultiInferMeta
param : [x_shape]
kernel :
func : einsum_grad
- backward_api : elementwise_pow_grad
forward : elementwise_pow(Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis=-1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : elementwise_pow_grad
- backward_api : elu_double_grad
forward : elu_grad (Tensor x, Tensor out, Tensor grad_out, float alpha)-> Tensor(grad_x)
args : (Tensor x, Tensor grad_out, Tensor grad_x_grad, float alpha)
output : Tensor(x_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, x]
kernel :
func : elu_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : elu_grad
forward : elu (Tensor x, float alpha) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, float alpha)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : elu_grad
backward : elu_double_grad
inplace : (out_grad -> x_grad)
- backward_api : embedding_grad
forward : embedding (Tensor x, Tensor weight, int64_t padding_idx=-1, bool sparse=false) -> Tensor(out)
args : (Tensor x, Tensor weight, Tensor out_grad, int64_t padding_idx=-1, bool sparse=false)
output : Tensor(weight_grad)
invoke : embedding_grad_impl(x, weight, out_grad, padding_idx, sparse, weight_grad)
- backward_api : erf_grad
forward : erf (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : erf_grad
data_type : out_grad
- backward_api : erfinv_grad
forward : erfinv (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : erfinv_grad
- backward_api : exp_grad
forward : exp (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : exp_grad
inplace : (out_grad -> x_grad)
- backward_api : expand_as_grad
forward : expand_as (Tensor x, Tensor y, int[] target_shape) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int[] target_shape)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : expand_as_grad
no_need_buffer : x
- backward_api : expand_double_grad
forward : expand_grad (Tensor x, Tensor grad_out, IntArray shape) -> Tensor(grad_x)
args : (Tensor grad_x_grad, IntArray shape)
output : Tensor(grad_out_grad)
infer_meta :
func : ExpandInferMeta
kernel :
func : expand
- backward_api : expand_grad
forward : expand (Tensor x, IntArray shape) -> Tensor(out)
args : (Tensor x, Tensor out_grad, IntArray shape)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : expand_grad
no_need_buffer : x
backward : expand_double_grad
- backward_api : expm1_grad
forward : expm1 (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : expm1_grad
inplace : (out_grad -> x_grad)
- backward_api : flatten_grad
forward : flatten(Tensor x, int start_axis, int stop_axis) -> Tensor(out), Tensor(xshape)
args : (Tensor xshape, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : KernelWithXShapeInferMeta
param : [xshape]
kernel :
func : flatten_grad
data_type: out_grad
backend: out_grad
layout: out_grad
inplace : (out_grad -> x_grad)
- backward_api : flip_grad
forward : flip (Tensor x, int[] axis) -> Tensor(out)
args : (Tensor out_grad, int[] axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : flip
- backward_api : floor_grad
forward : floor(Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : floor_grad
inplace : (out_grad -> x_grad)
- backward_api : fmax_grad
forward : fmax(Tensor x, Tensor y, int axis) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : fmax_grad
- backward_api : fmin_grad
forward : fmin(Tensor x, Tensor y, int axis) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : fmin_grad
- backward_api : frobenius_norm_grad
forward : frobenius_norm(Tensor x, int64_t[] axis, bool keep_dim, bool reduce_all) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis, bool keep_dim, bool reduce_all)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : frobenius_norm_grad
- backward_api : gather_grad
forward : gather(Tensor x, Tensor index, Scalar axis=0) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad, Scalar axis=0, bool overwrite=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
data_type: x
func : gather_grad
no_need_buffer : x
- backward_api : gather_nd_grad
forward : gather_nd (Tensor x, Tensor index) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : gather_nd_grad
no_need_buffer : x
- backward_api : gelu_grad
forward : gelu(Tensor x, bool approximate) -> Tensor(out)
args : (Tensor x, Tensor out_grad, bool approximate)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : gelu_grad
- backward_api : graph_send_recv_grad
forward : graph_send_recv (Tensor x, Tensor src_index, Tensor dst_index, str pool_type = "SUM", int64_t out_size = 0) -> Tensor(out), Tensor(dst_count)
args : (Tensor x, Tensor src_index, Tensor dst_index, Tensor out, Tensor dst_count, Tensor out_grad, str pool_type = "SUM")
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param : [x]
kernel :
func : graph_send_recv_grad
data_type : out_grad
optional: out, dst_count
- backward_api : group_norm_grad
forward : group_norm (Tensor x, Tensor scale, Tensor bias, float epsilon, int groups, str data_layout) -> Tensor(y), Tensor(mean), Tensor(variance)
args : (Tensor x, Tensor scale, Tensor bias, Tensor y, Tensor mean, Tensor variance, Tensor y_grad, float epsilon, int groups, str data_layout)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [y, scale, bias]
kernel :
func : group_norm_grad
data_type : y_grad
optional: scale, bias
inplace : (y_grad -> x_grad)
- backward_api : gumbel_softmax_grad
forward : gumbel_softmax (Tensor x, float temperature, bool hard, int axis) -> Tensor(out)
args : (Tensor out, Tensor out_grad, int axis)
output : Tensor(x_grad)
infer_meta :
func : GumbelSoftmaxGradInferMeta
param : [out, out_grad, axis]
kernel :
func : gumbel_softmax_grad
- backward_api : hard_shrink_grad
forward : hard_shrink (Tensor x, float threshold) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float threshold)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_shrink_grad
inplace : (out_grad -> x_grad)
- backward_api : hard_sigmoid_grad
forward : hard_sigmoid (Tensor x, float slope, float offset) -> Tensor(out)
args : (Tensor out, Tensor out_grad, float slope, float offset)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : hard_sigmoid_grad
inplace : (out_grad -> x_grad)
- backward_api : hard_swish_grad
forward : hard_swish (Tensor x, float threshold = 6.0, float scale = 6.0, float offset = 3.0) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float threshold, float scale, float offset)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : hard_swish_grad
inplace : (out_grad -> x_grad)
- backward_api : huber_loss_grad
forward : huber_loss (Tensor input, Tensor label, float delta) -> Tensor(out), Tensor(residual)
args : (Tensor residual, Tensor out_grad, float delta)
output : Tensor(input_grad), Tensor(label_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [residual, residual]
kernel :
func : huber_loss_grad
- backward_api : imag_grad
forward : imag (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
invoke : imag_grad_impl(out_grad, x_grad)
- backward_api : index_sample_grad
forward : index_sample (Tensor x, Tensor index) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : index_sample_grad
data_type : out_grad
no_need_buffer : x
- backward_api : index_select_grad
forward : index_select(Tensor x, Tensor index, int dim) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad, int dim)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : index_select_grad
data_type : x
no_need_buffer : x
- backward_api : instance_norm_double_grad
forward : instance_norm_grad(Tensor x, Tensor fwd_scale, Tensor saved_mean, Tensor saved_variance, Tensor grad_y, float epsilon) -> Tensor(grad_x), Tensor(grad_scale), Tensor(grad_bias)
args : (Tensor x, Tensor fwd_scale, Tensor saved_mean, Tensor saved_variance, Tensor grad_y, Tensor grad_x_grad, Tensor grad_scale_grad, Tensor grad_bias_grad, float epsilon)
output : Tensor(x_grad), Tensor(fwd_scale_grad), Tensor(grad_y_grad)
infer_meta :
func : InstanceNormDoubleGradInferMeta
kernel :
func : instance_norm_double_grad
data_type : x
optional : fwd_scale, grad_x_grad, grad_scale_grad, grad_bias_grad
- backward_api : instance_norm_grad
forward : instance_norm(Tensor x, Tensor scale, Tensor bias, float epsilon) -> Tensor(y), Tensor(saved_mean), Tensor(saved_variance)
args : (Tensor x, Tensor scale, Tensor saved_mean, Tensor saved_variance, Tensor y_grad, float epsilon)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
infer_meta :
func : InstanceNormGradInferMeta
kernel :
func : instance_norm_grad
data_type : x
optional : scale
backward : instance_norm_double_grad
- backward_api : kldiv_loss_grad
forward : kldiv_loss(Tensor x, Tensor label, str reduction) -> Tensor(out)
args : (Tensor x, Tensor label, Tensor out_grad, str reduction)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : kldiv_loss_grad
no_need_buffer : x
- backward_api : kron_grad
forward : kron (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : kron_grad
data_type : out_grad
- backward_api : kthvalue_grad
forward : kthvalue(Tensor x, int k, int axis, bool keepdim) -> Tensor(out), Tensor(indices)
args : (Tensor x, Tensor indices, Tensor out_grad, int k, int axis, bool keepdim)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : kthvalue_grad
- backward_api : label_smooth_grad
forward : label_smooth (Tensor label, Tensor prior_dist, float epsilon) -> Tensor(out)
args : (Tensor out_grad, float epsilon)
output : Tensor(label_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : label_smooth_grad
- backward_api : layer_norm_grad
forward : layer_norm (Tensor x, Tensor scale, Tensor bias, float epsilon, int begin_norm_axis, bool is_test) -> Tensor(out), Tensor(mean), Tensor(variance)
args : (Tensor x, Tensor scale, Tensor bias, Tensor mean, Tensor variance, Tensor out_grad, float epsilon, int begin_norm_axis, bool is_test)
output : Tensor(x_grad), Tensor(scale_grad), Tensor(bias_grad)
infer_meta :
func : LayerNormGradInferMeta
param : [x, scale, bias]
kernel :
func : layer_norm_grad
data_type : out_grad
no_need_buffer : bias
optional : scale, bias
- backward_api : leaky_relu_double_grad
forward : leaky_relu_grad (Tensor x, Tensor grad_out, float alpha) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_x_grad, float alpha)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [grad_x_grad]
kernel :
func : leaky_relu_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : leaky_relu_grad
forward : leaky_relu (Tensor x, float alpha) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float alpha)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : leaky_relu_grad
backward : leaky_relu_double_grad
inplace : (out_grad -> x_grad)
- backward_api : lerp_grad
forward : lerp (Tensor x, Tensor y, Tensor weight) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor weight, Tensor out, Tensor out_grad)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : lerp_grad
- backward_api : lgamma_grad
forward : lgamma(Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : lgamma_grad
- backward_api : log10_grad
forward : log10 (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : log10_grad
inplace : (out_grad -> x_grad)
- backward_api : log1p_grad
forward : log1p (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : log1p_grad
inplace : (out_grad -> x_grad)
- backward_api : log2_grad
forward : log2 (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : log2_grad
inplace : (out_grad -> x_grad)
- backward_api : log_double_grad
forward : log_grad (Tensor x, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_out, Tensor grad_x_grad)
output : Tensor(x_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, x]
kernel :
func : log_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : log_grad
forward : log (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : log_grad
backward : log_double_grad
inplace : (out_grad -> x_grad)
- backward_api : log_loss_grad
forward : log_loss (Tensor input, Tensor label, float epsilon) -> Tensor(out)
args : (Tensor input, Tensor label, Tensor out_grad, float epsilon)
output : Tensor(input_grad)
infer_meta :
func : UnchangedInferMeta
param : [input]
kernel :
func : log_loss_grad
- backward_api : log_softmax_grad
forward : log_softmax(Tensor x, int axis) -> Tensor(out)
args : (Tensor out, Tensor out_grad, int axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out]
kernel :
func : log_softmax_grad
- backward_api : logcumsumexp_grad
forward : logcumsumexp(Tensor x, int axis, bool flatten, bool exclusive, bool reverse) -> Tensor(out)
infer_meta :
func : UnchangedInferMeta
param : [x]
args : (Tensor x, Tensor out, Tensor out_grad, int axis, bool flatten, bool exclusive, bool reverse)
output : Tensor(x_grad)
kernel :
func : logcumsumexp_grad
- backward_api : logit_grad
forward : logit (Tensor x, float eps = 1e-6f) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float eps)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : logit_grad
- backward_api : logsigmoid_grad
forward : logsigmoid (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : logsigmoid_grad
inplace : (out_grad -> x_grad)
- backward_api : logsumexp_grad
forward : logsumexp(Tensor x, int64_t[] axis, bool keepdim, bool reduce_all) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis, bool keepdim, bool reduce_all)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : logsumexp_grad
- backward_api : masked_select_grad
forward : masked_select (Tensor x, Tensor mask) -> Tensor(out)
args : (Tensor x, Tensor mask, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : masked_select_grad
data_type : x
no_need_buffer : x
- backward_api : matmul_double_grad
forward : matmul_grad (Tensor x, Tensor y, Tensor grad_out, bool transpose_x=false, bool transpose_y=false) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor x, Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, bool transpose_x=false, bool transpose_y=false)
output : Tensor(x_grad), Tensor(y_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [x, y, grad_out]
kernel :
func : matmul_double_grad
backward : matmul_triple_grad
optional : grad_x_grad, grad_y_grad
- backward_api : matmul_grad
forward : matmul (Tensor x, Tensor y, bool transpose_x=false, bool transpose_y=false) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, bool transpose_x=false, bool transpose_y=false)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : matmul_grad
backward : matmul_double_grad
- backward_api : matmul_triple_grad
forward : matmul_double_grad (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, bool transpose_x=false, bool transpose_y=false) -> Tensor(grad_x), Tensor(grad_y), Tensor(grad_grad_out)
args : (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, Tensor grad_x_grad, Tensor grad_y_grad, Tensor grad_grad_out_grad, bool transpose_x=false, bool transpose_y=false)
output : Tensor(x_grad), Tensor(y_grad), Tensor(fwd_grad_out_grad), Tensor(fwd_grad_grad_x_grad), Tensor(fwd_grad_grad_y_grad)
infer_meta :
func : GeneralQuinaryGradInferMeta
param : [x, y, fwd_grad_out, fwd_grad_grad_x, fwd_grad_grad_y]
kernel :
func : matmul_triple_grad
optional : grad_x_grad, grad_y_grad, grad_grad_out_grad
- backward_api : matrix_power_grad
forward : matrix_power (Tensor x, int n) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int n)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : matrix_power_grad
- backward_api : max_grad
forward: max (Tensor x, int64_t[] dims={}, bool keep_dim=false) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : max_grad
- backward_api : max_pool2d_with_index_grad
forward : max_pool2d_with_index(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive) -> Tensor(out), Tensor(mask)
args : (Tensor x, Tensor mask, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
output : Tensor(x_grad)
infer_meta :
func : MaxPoolWithIndexGradInferMeta
kernel :
func : max_pool2d_with_index_grad
- backward_api : max_pool3d_with_index_grad
forward : max_pool3d_with_index(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive) -> Tensor(out), Tensor(mask)
args : (Tensor x, Tensor mask, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool global_pooling, bool adaptive)
output : Tensor(x_grad)
infer_meta :
func : MaxPoolWithIndexGradInferMeta
kernel :
func : max_pool3d_with_index_grad
- backward_api : maximum_grad
forward : maximum(Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis=-1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : maximum_grad
- backward_api : maxout_grad
forward : maxout(Tensor x, int groups, int axis) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int groups, int axis)
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param: [x]
kernel :
func : maxout_grad
- backward_api : mean_all_grad
forward : mean_all(Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : mean_all_grad
- backward_api : mean_double_grad
forward: mean_grad (Tensor x, Tensor grad_out, int64_t[] dims={}, bool keep_dim=false, bool reduce_all = false) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(grad_out_grad)
invoke : mean(grad_x_grad, dims, keep_dim)
- backward_api : mean_grad
forward: mean (Tensor x, int64_t[] dims={}, bool keep_dim=false) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : mean_grad
backward : mean_double_grad
no_need_buffer : x
- backward_api : meshgrid_grad
forward : meshgrid (Tensor[] inputs) -> Tensor[](outputs)
args : (Tensor[] inputs, Tensor[] outputs_grad)
output : Tensor[](inputs_grad){inputs.size()}
infer_meta :
func : MeshgridGradInferMeta
kernel :
func : meshgrid_grad
- backward_api : min_grad
forward: min (Tensor x, int64_t[] dims={}, bool keep_dim=false) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : min_grad
- backward_api : minimum_grad
forward : minimum(Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis=-1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, y]
kernel :
func : minimum_grad
- backward_api : mish_grad
forward : mish (Tensor x, float threshold) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float threshold)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : mish_grad
inplace : (out_grad -> x_grad)
- backward_api : mode_grad
forward : mode(Tensor x, int axis, bool keepdim) -> Tensor(out), Tensor(indices)
args : (Tensor x, Tensor indices, Tensor out_grad, int axis, bool keepdim)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : mode_grad
- backward_api : modulo_grad
forward : modulo (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : modulo_grad
no_need_buffer : x, y
- backward_api : multi_dot_grad
forward : multi_dot (Tensor[] x) -> Tensor(out)
args : (Tensor[] x, Tensor out_grad)
output : Tensor[](x_grad) {x.size()}
infer_meta :
func : MultiDotGradInferMeta
kernel :
func : multi_dot_grad
- backward_api : multiplex_grad
forward : multiplex (Tensor[] ins, Tensor ids) -> Tensor(out)
args : (Tensor[] ins, Tensor ids, Tensor out_grad)
output : Tensor[](ins_grad){ins.size()}
infer_meta :
func : MultiplexGradInferMeta
param : [ids, out_grad]
kernel :
func : multiplex_grad
param : [ids, out_grad]
- backward_api : multiply_double_grad
forward : multiply_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor x, Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [x, y, grad_out]
kernel :
func : multiply_double_grad
optional : grad_x_grad, grad_y_grad
backward : multiply_triple_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : multiply_grad
forward : multiply (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : multiply_grad
backward : multiply_double_grad
- backward_api : multiply_triple_grad
forward : multiply_double_grad (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, int aixs = -1) -> Tensor(grad_x), Tensor(grad_y), Tensor(grad_grad_out)
args : (Tensor x, Tensor y, Tensor fwd_grad_out, Tensor fwd_grad_grad_x, Tensor fwd_grad_grad_y, Tensor grad_x_grad, Tensor grad_y_grad, Tensor grad_grad_out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad), Tensor(fwd_grad_out_grad), Tensor(fwd_grad_grad_x_grad), Tensor(fwd_grad_grad_y_grad)
infer_meta :
func : GeneralQuinaryGradInferMeta
param : [x, y, fwd_grad_out, x, y]
kernel :
func : multiply_triple_grad
optional : fwd_grad_grad_x, fwd_grad_grad_y, grad_grad_out_grad
- backward_api : mv_grad
forward : mv (Tensor x, Tensor vec) -> Tensor(out)
args : (Tensor x, Tensor vec, Tensor out_grad)
output : Tensor(x_grad), Tensor(vec_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, vec]
kernel :
func : mv_grad
- backward_api : nll_loss_grad
forward : nll_loss (Tensor input, Tensor label, Tensor weight, int64_t ignore_index, str reduction) -> Tensor(out), Tensor(total_weight)
args : (Tensor input, Tensor label, Tensor weight, Tensor total_weight, Tensor out_grad, int64_t ignore_index, str reduction)
output : Tensor(input_grad)
infer_meta :
func : NllLossGradInferMeta
kernel :
func : nll_loss_grad
data_type : input
optional : weight
- backward_api : norm_grad
forward : norm (Tensor x, int axis, float epsilon, bool is_test) -> Tensor(out), Tensor(norm)
args : (Tensor x, Tensor norm, Tensor out_grad, int axis, float epsilon, bool is_test)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : norm_grad
- backward_api : p_norm_grad
forward : p_norm(Tensor x, float porder, int axis, float epsilon, bool keepdim, bool asvector=false) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, float porder, int axis, float epsilon, bool keepdim, bool asvector)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : p_norm_grad
- backward_api : pad3d_double_grad
forward : pad3d_grad(Tensor x, Tensor grad_out, IntArray paddings, str mode, float pad_value, str data_format) -> Tensor(grad_x)
args : (Tensor grad_x_grad, IntArray paddings, str mode, float pad_value, str data_format)
output : Tensor(grad_out_grad)
infer_meta :
func : Pad3dInferMeta
kernel :
func : pad3d
- backward_api : pad3d_grad
forward : pad3d(Tensor x, IntArray paddings, str mode, float pad_value, str data_format) -> Tensor(out)
args : (Tensor x, Tensor out_grad, IntArray paddings, str mode, float pad_value, str data_format)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : pad3d_grad
no_need_buffer : x
backward : pad3d_double_grad
- backward_api : pad_double_grad
forward : pad_grad(Tensor x, Tensor grad_out, int[] paddings, float pad_value) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int[] paddings, float pad_value)
output : Tensor(grad_out_grad)
infer_meta :
func : PadInferMeta
kernel :
func : pad
- backward_api : pad_grad
forward : pad(Tensor x, int[] paddings, float pad_value) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int[] paddings, float pad_value)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : pad_grad
param: [out_grad, paddings, pad_value]
no_need_buffer : x
backward : pad_double_grad
- backward_api : pixel_shuffle_grad
forward : pixel_shuffle (Tensor x, int upscale_factor, str data_format) -> Tensor(out)
args : (Tensor out_grad, int upscale_factor, str data_format)
output : Tensor(x_grad)
infer_meta :
func : PixelShuffleGradInferMeta
kernel :
func : pixel_shuffle_grad
- backward_api : poisson_grad
forward : poisson (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : poisson_grad
- backward_api : pool2d_double_grad
forward : pool2d_grad(Tensor x, Tensor out, Tensor grad_out, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(grad_out_grad)
infer_meta :
func : PoolInferMeta
kernel :
func : pool2d_double_grad
use_gpudnn : true
- backward_api : pool2d_grad
forward : pool2d(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(x_grad)
infer_meta :
func : PoolGradInferMeta
kernel :
func : pool2d_grad
use_gpudnn : true
backward : pool2d_double_grad
- backward_api : pool2d_grad_gpudnn_unused
forward : pool2d_gpudnn_unused(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(x_grad)
infer_meta :
func : PoolGradInferMeta
kernel :
func : pool2d_grad
use_gpudnn : false
- backward_api : pool3d_grad
forward : pool3d(Tensor x, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int[] kernel_size, int[] strides, int[] paddings, bool ceil_mode, bool exclusive, str data_format, str pooling_type, bool global_pooling, bool adaptive, str padding_algorithm)
output : Tensor(x_grad)
infer_meta :
func : PoolGradInferMeta
kernel :
func : pool3d_grad
use_gpudnn : true
- backward_api : pow_grad
forward : pow(Tensor x, Scalar s) -> Tensor(out)
args : (Tensor x, Tensor out_grad, Scalar s=-1)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [x]
kernel :
func : pow_grad
inplace : (out_grad -> x_grad)
- backward_api : prelu_grad
forward : prelu(Tensor x, Tensor alpha, str data_format, str mode) -> Tensor(out)
args : (Tensor x, Tensor alpha, Tensor out_grad, str data_format, str mode)
output : Tensor(x_grad), Tensor(alpha_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param: [x, alpha]
kernel :
func : prelu_grad
- backward_api : psroi_pool_grad
forward : psroi_pool (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, int output_channels, float spatial_scale) -> Tensor(out)
args : (Tensor x, Tensor boxes, Tensor boxes_num, Tensor out_grad, int pooled_height, int pooled_width, int output_channels, float spatial_scale)
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param : [x]
kernel :
func : psroi_pool_grad
data_type : x
optional : boxes_num
# output is optional
- backward_api : put_along_axis_grad
forward : put_along_axis (Tensor x, Tensor index, Tensor value, int axis, str reduce) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad, int axis, str reduce)
output : Tensor(x_grad), Tensor(value_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, index]
kernel :
func : put_along_axis_grad
- backward_api : real_grad
forward : real (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
invoke : real_grad_impl(out_grad, x_grad)
- backward_api : reciprocal_grad
forward : reciprocal (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : reciprocal_grad
inplace : (out_grad -> x_grad)
- backward_api : reduce_prod_grad
forward : reduce_prod (Tensor x, int64_t[] dims, bool keep_dim, bool reduce_all) -> Tensor(out)
args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] dims, bool keep_dim, bool reduce_all)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : prod_grad
- backward_api : relu_double_grad
forward : relu_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor grad_x_grad)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : relu_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : relu_grad
forward : relu (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : relu_grad
backward: relu_double_grad
inplace : (out_grad -> x_grad)
- backward_api : reshape_double_grad
forward : reshape_grad (Tensor xshape, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor grad_out, Tensor grad_x_grad)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [grad_out]
kernel :
func : reshape_double_grad
no_need_buffer : grad_out
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : reshape_grad
forward : reshape (Tensor x, IntArray shape) -> Tensor(out), Tensor(xshape)
args : (Tensor xshape, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : KernelWithXShapeInferMeta
param : [xshape]
kernel :
func : reshape_grad
param : [out_grad]
data_type: out_grad
backend: out_grad
layout: out_grad
backward : reshape_double_grad
inplace : (out_grad -> x_grad)
- backward_api : roi_align_grad
forward : roi_align (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale, int sampling_ratio, bool aligned) -> Tensor(out)
args : (Tensor x, Tensor boxes, Tensor boxes_num, Tensor out_grad, int pooled_height, int pooled_width, float spatial_scale, int sampling_ratio, bool aligned)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : roi_align_grad
data_type : boxes
no_need_buffer : x
optional : boxes_num
- backward_api : roi_pool_grad
forward : roi_pool (Tensor x, Tensor boxes, Tensor boxes_num, int pooled_height, int pooled_width, float spatial_scale) -> Tensor(out), Tensor(arg_max)
args : (Tensor x, Tensor boxes, Tensor boxes_num, Tensor arg_max, Tensor out_grad, int pooled_height, int pooled_width, float spatial_scale)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : roi_pool_grad
data_type : x
optional : boxes_num
- backward_api : roll_grad
forward : roll(Tensor x, IntArray shifts, int64_t[] axis) -> Tensor(out)
args : (Tensor x, Tensor out_grad, IntArray shifts, int64_t[] axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : roll_grad
data_type : x
no_need_buffer : x
- backward_api : round_grad
forward : round(Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param: [out_grad]
kernel :
func : round_grad
inplace : (out_grad -> x_grad)
- backward_api : rsqrt_double_grad
forward : rsqrt_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor grad_x, Tensor grad_x_grad)
output : Tensor(out_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [out, out]
kernel :
func : rsqrt_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : rsqrt_grad
forward : rsqrt (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : rsqrt_grad
backward : rsqrt_double_grad
inplace : (out_grad -> x_grad)
- backward_api : scale_double_grad
forward : scale_grad (Tensor grad_out, Scalar scale, float bias, bool bias_after_scale) -> Tensor(grad_x)
args : (Tensor grad_x_grad, Scalar scale=1.0, float bias=0.0, bool bias_after_scale=true)
output : Tensor(grad_out_grad)
invoke : scale(grad_x_grad, scale, 0.0, bias_after_scale)
backward : scale_triple_grad
- backward_api : scale_grad
forward : scale (Tensor x, Scalar scale, float bias, bool bias_after_scale) -> Tensor(out)
args : (Tensor out_grad, Scalar scale=1.0, float bias=0.0, bool bias_after_scale=true)
output : Tensor(x_grad)
invoke : scale(out_grad, scale, 0.0, bias_after_scale)
backward : scale_double_grad
inplace : (out_grad -> x_grad)
- backward_api : scale_triple_grad
forward : scale_double_grad (Tensor grad_grad_x, Scalar scale, float bias, bool bias_after_scale) -> Tensor(grad_grad_out)
args : (Tensor grad_grad_out_grad, Scalar scale=1.0, float bias=0.0, bool bias_after_scale=true)
output : Tensor(grad_grad_x_grad)
invoke : scale(grad_grad_out_grad, scale, 0.0, bias_after_scale)
- backward_api : scatter_grad
forward : scatter (Tensor x, Tensor index, Tensor updates, bool overwrite) -> Tensor(out)
args : (Tensor index, Tensor updates, Tensor out_grad, bool overwrite)
output : Tensor(x_grad), Tensor(updates_grad)
infer_meta :
func : ScatterGradInferMeta
param : [index, updates, out_grad, overwrite]
kernel :
func : scatter_grad
no_need_buffer : updates
- backward_api : scatter_nd_add_grad
forward : scatter_nd_add (Tensor x, Tensor index, Tensor updates) -> Tensor(out)
args : (Tensor index, Tensor updates, Tensor out_grad)
output : Tensor(x_grad), Tensor(updates_grad)
infer_meta :
func : ScatterNdAddGradInferMeta
param : [index, updates, out_grad]
kernel :
func : scatter_nd_add_grad
no_need_buffer : updates
- backward_api : segment_pool_grad
forward : segment_pool (Tensor x, Tensor segment_ids, str pooltype) -> Tensor(out), Tensor(summed_ids)
args : (Tensor x, Tensor segment_ids, Tensor out, Tensor summed_ids, Tensor out_grad, str pooltype)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : segment_pool_grad
data_type : x
optional : summed_ids
- backward_api : selu_grad
forward : selu (Tensor x, float scale, float alpha) -> Tensor(out)
args : (Tensor out, Tensor out_grad, float scale, float alpha)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : selu_grad
- backward_api : sigmoid_cross_entropy_with_logits_grad
forward : sigmoid_cross_entropy_with_logits (Tensor x, Tensor label, bool normalize, int ignore_index) -> Tensor(out)
args : (Tensor x, Tensor label, Tensor out_grad, bool normalize, int ignore_index)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : sigmoid_cross_entropy_with_logits_grad
inplace : (out_grad -> x_grad)
- backward_api : sigmoid_double_grad
forward : sigmoid_grad (Tensor out, Tensor fwd_grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor fwd_grad_out, Tensor grad_x_grad)
output : Tensor(out_grad), Tensor(fwd_grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [out, fwd_grad_out]
kernel :
func : sigmoid_double_grad
backward : sigmoid_triple_grad
inplace : (grad_x_grad -> fwd_grad_out_grad)
- backward_api : sigmoid_grad
forward : sigmoid (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : sigmoid_grad
backward : sigmoid_double_grad
inplace : (out_grad -> x_grad)
- backward_api : sigmoid_triple_grad
forward : sigmoid_double_grad (Tensor out, Tensor fwd_grad_out, Tensor grad_grad_x) -> Tensor(grad_out), Tensor(grad_grad_out)
args : (Tensor out, Tensor fwd_grad_out, Tensor grad_grad_x, Tensor grad_out_grad, Tensor grad_grad_out_grad)
output : Tensor(out_grad), Tensor(fwd_grad_out_grad), Tensor(grad_grad_x_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [out, fwd_grad_out, grad_grad_x]
kernel :
func : sigmoid_triple_grad
optional : grad_grad_out_grad
inplace : (grad_grad_x -> fwd_grad_out_grad)
- backward_api : silu_grad
forward : silu (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : silu_grad
inplace : (out_grad -> x_grad)
- backward_api : sin_grad
forward : sin (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : sin_grad
inplace : (out_grad -> x_grad)
- backward_api : sinh_grad
forward : sinh (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : sinh_grad
inplace : (out_grad -> x_grad)
- backward_api : slice_grad
forward : slice (Tensor input, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis) -> Tensor(out)
args : (Tensor input, Tensor out_grad, int64_t[] axes, IntArray starts, IntArray ends, int64_t[] infer_flags, int64_t[] decrease_axis)
output : Tensor(input_grad)
infer_meta :
func : UnchangedInferMeta
param : [input]
kernel :
func : slice_grad
no_need_buffer : input
- backward_api : soft_shrink_grad
forward : soft_shrink (Tensor x, float lambda) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float lambda)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : soft_shrink_grad
inplace : (out_grad -> x_grad)
- backward_api : softmax_grad
forward : softmax (Tensor x, int axis) -> Tensor(out)
args : (Tensor out, Tensor out_grad, int axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : softmax_grad
use_gpudnn : true
- backward_api : split_grad
forward : split (Tensor x, IntArray num_or_sections, Scalar axis) -> Tensor[](out)
args : (Tensor[] out_grad, Scalar axis = -1)
output : Tensor(x_grad)
invoke : concat( out_grad, axis)
# TODO(zhangyunfei) The config of double grad and triple grad will be supported in the future.
- backward_api : sqrt_double_grad
forward : sqrt_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor grad_x, Tensor grad_x_grad)
output : Tensor(out_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [out, out]
kernel :
func : sqrt_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : sqrt_grad
forward : sqrt (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : sqrt_grad
backward : sqrt_double_grad
inplace : (out_grad -> x_grad)
- backward_api : square_double_grad
forward : square_grad (Tensor x, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor x, Tensor grad_out, Tensor grad_x_grad)
output : Tensor(x_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, x]
kernel :
func : square_double_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : square_grad
forward : square (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : square_grad
backward : square_double_grad
inplace : (out_grad -> x_grad)
- backward_api : squeeze_double_grad
forward : squeeze_grad(Tensor xshape, Tensor grad_out, int[] axes) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int[] axes)
output : Tensor(grad_out_grad)
invoke: squeeze(grad_x_grad, axes)
- backward_api : squeeze_grad
forward : squeeze(Tensor x, int[] axes) -> Tensor(out), Tensor(xshape)
args : (Tensor xshape, Tensor out_grad, int[] axes)
output : Tensor(x_grad)
infer_meta :
func : KernelWithXShapeInferMeta
param: [xshape]
kernel :
func : squeeze_grad
inplace : (out_grad -> x_grad)
backward: squeeze_double_grad
- backward_api : stack_grad
forward : stack (Tensor[] x, int axis) -> Tensor(out)
args : (Tensor[] x, Tensor out_grad, int axis)
output : Tensor[](x_grad){x.size()}
infer_meta :
func : StackGradInferMeta
param: [out_grad, axis]
kernel :
func : stack_grad
param : [out_grad, axis]
no_need_buffer : x
- backward_api : strided_slice_grad
forward : strided_slice (Tensor x, int[] axes, IntArray starts, IntArray ends, IntArray strides) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int[] axes, IntArray starts, IntArray ends, IntArray strides)
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param : [x]
kernel :
func : strided_slice_grad
no_need_buffer : x
- backward_api : subtract_double_grad
forward : subtract_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y)
args : (Tensor y, Tensor grad_out, Tensor grad_x_grad, Tensor grad_y_grad, int axis = -1)
output : Tensor(grad_out_grad)
infer_meta :
func : UnchangedInferMeta
param : [grad_out]
kernel :
func : subtract_double_grad
optional : grad_x_grad, grad_y_grad
no_need_buffer : y, grad_out
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : subtract_grad
forward : subtract (Tensor x, Tensor y) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out_grad, int axis = -1)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : subtract_grad
no_need_buffer : x, y
backward : subtract_double_grad
inplace : (out_grad -> x_grad)
- backward_api : sum_double_grad
forward : sum_grad (Tensor x, Tensor grad_out, int64_t[] dims, bool keep_dim, bool reduce_all=false) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int64_t[] dims={}, bool keep_dim=false)
output : Tensor(grad_out_grad)
invoke : sum(grad_x_grad, dims, grad_x_grad.dtype(), keep_dim)
backward : sum_triple_grad
- backward_api : sum_grad
forward : sum (Tensor x, int64_t[] dims={}, DataType out_dtype=DataType::UNDEFINED, bool keep_dim=false) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int64_t[] dims, bool keep_dim, bool reduce_all=false)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : sum_grad
no_need_buffer : x
backward : sum_double_grad
- backward_api : sum_triple_grad
forward : sum_double_grad (Tensor grad_grad_x, int64_t[] dims={}, bool keep_dim=false) -> Tensor(grad_grad_out)
args : (Tensor grad_grad_x, Tensor grad_grad_out_grad, int64_t[] dims={}, bool keep_dim=false, bool reduce_all=false)
output : Tensor(grad_grad_x_grad)
invoke : sum_grad(grad_grad_x, grad_grad_out_grad, dims, keep_dim, reduce_all, grad_grad_x_grad)
- backward_api : swish_grad
forward : swish (Tensor x, float beta=1.0) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float bete=1.0)
output : Tensor(x_grad)
infer_meta :
func : GeneralUnaryGradInferMeta
param : [x]
kernel :
func : swish_grad
inplace : (out_grad -> x_grad)
- backward_api : take_along_axis_grad
forward : take_along_axis (Tensor x, Tensor index, int axis) -> Tensor(out)
args : (Tensor x, Tensor index, Tensor out_grad, int axis)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : take_along_axis_grad
- backward_api : tan_grad
forward : tan (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : tan_grad
inplace : (out_grad -> x_grad)
- backward_api : tanh_double_grad
forward : tanh_grad (Tensor out, Tensor grad_out) -> Tensor(grad_x)
args : (Tensor out, Tensor grad_out, Tensor grad_x_grad)
output : Tensor(out_grad), Tensor(grad_out_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [out, out]
kernel :
func : tanh_double_grad
backward : tanh_triple_grad
inplace : (grad_x_grad -> grad_out_grad)
- backward_api : tanh_grad
forward : tanh (Tensor x) -> Tensor(out)
args : (Tensor out, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out]
kernel :
func : tanh_grad
backward : tanh_double_grad
inplace : (out_grad -> x_grad)
- backward_api : tanh_shrink_grad
forward : tanh_shrink (Tensor x) -> Tensor(out)
args : (Tensor x, Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : tanh_shrink_grad
inplace : (out_grad -> x_grad)
- backward_api : tanh_triple_grad
forward : tanh_double_grad (Tensor out, Tensor grad_out_forward, Tensor grad_x_grad_forward) -> Tensor(grad_out_new), Tensor(grad_out_grad)
args : (Tensor out, Tensor grad_out_forward, Tensor grad_x_grad_forward, Tensor grad_out_new_grad, Tensor grad_out_grad_grad)
output : Tensor(out_grad), Tensor(grad_out_forward_grad), Tensor(grad_x_grad_forward_grad)
infer_meta :
func : GeneralTernaryGradInferMeta
param : [out, out, grad_x_grad_forward]
kernel :
func : tanh_triple_grad
inplace : (grad_x_grad_forward -> grad_out_forward_grad)
- backward_api : thresholded_relu_grad
forward : thresholded_relu (Tensor x, float threshold) -> Tensor(out)
args : (Tensor x, Tensor out_grad, float threshold)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : thresholded_relu_grad
inplace : (out_grad -> x_grad)
- backward_api : tile_double_grad
forward : tile_grad (Tensor x, Tensor grad_out, IntArray repeat_times) -> Tensor(grad_x)
args : (Tensor grad_x_grad, IntArray repeat_times)
output : Tensor(grad_out_grad)
infer_meta :
func : TileInferMeta
kernel :
func : tile
- backward_api : tile_grad
forward : tile (Tensor x, IntArray repeat_times) -> Tensor(out)
args : (Tensor x, Tensor out_grad, IntArray repeat_times)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : tile_grad
no_need_buffer : x
backward : tile_double_grad
- backward_api : top_k_grad
forward : top_k (Tensor x, Scalar k, int axis = -1, bool largest = true, bool sorted = true) -> Tensor(out), Tensor(indices)
args : (Tensor x, Tensor indices, Tensor out_grad, Scalar k = -1, int axis = -1, bool largest = true, bool sorted = true)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : top_k_grad
- backward_api : trace_grad
forward : trace (Tensor x, int offset, int axis1, int axis2) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int offset, int axis1, int axis2)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : trace_grad
no_need_buffer : x
- backward_api : transpose_double_grad
forward : transpose_grad (Tensor grad_out, int[] axis) -> Tensor(grad_x)
args : (Tensor grad_x_grad, int[] axis)
output : Tensor(grad_out_grad)
invoke : transpose(grad_x_grad, axis)
- backward_api : transpose_grad
forward : transpose (Tensor x, int[] axis) -> Tensor(out)
args : (Tensor out_grad, int[] axis)
output : Tensor(x_grad)
infer_meta :
func : TransposeGradInferMeta
param : [out_grad, axis]
kernel :
func : transpose_grad
backward : transpose_double_grad
- backward_api : triangular_solve_grad
forward : triangular_solve (Tensor x, Tensor y, bool upper, bool tranpose, bool unitriangular) -> Tensor(out)
args : (Tensor x, Tensor y, Tensor out, Tensor out_grad, bool upper, bool tranpose, bool unitriangular)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : triangular_solve_grad
- backward_api : tril_triu_grad
forward : tril_triu(Tensor x, int diagonal, bool lower) -> Tensor(out)
args : (Tensor out_grad, int diagonal, bool lower)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : tril_triu_grad
- backward_api : trunc_grad
forward : trunc (Tensor x) -> Tensor(out)
args : (Tensor out_grad)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [out_grad]
kernel :
func : trunc_grad
- backward_api : unbind_grad
forward : unbind (Tensor input, int axis) -> Tensor[](out)
args : (Tensor[] out_grad, int axis)
output : Tensor(input_grad)
invoke : stack(out_grad, axis)
- backward_api : unfold_grad
forward : unfold (Tensor x, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations) -> Tensor(out)
args : (Tensor x, Tensor out_grad, int[] kernel_sizes, int[] strides, int[] paddings, int[] dilations)
output : Tensor(x_grad)
infer_meta :
func : UnchangedInferMeta
param : [x]
kernel :
func : unfold_grad
no_need_buffer : x
- backward_api : unsqueeze_double_grad
forward : unsqueeze_grad(Tensor xshape, Tensor grad_out, IntArray axes) -> Tensor(grad_x)
args : (Tensor grad_x_grad, IntArray axes)
output : Tensor(grad_out_grad)
invoke : unsqueeze(grad_x_grad, axes)
- backward_api : unsqueeze_grad
forward : unsqueeze(Tensor x, IntArray axes) -> Tensor(out), Tensor(xshape)
args : (Tensor xshape, Tensor out_grad, IntArray axes)
output : Tensor(x_grad)
infer_meta :
func : KernelWithXShapeInferMeta
param: [xshape]
kernel :
func : unsqueeze_grad
param: [xshape, out_grad]
inplace : (out_grad -> x_grad)
backward : unsqueeze_double_grad
- backward_api : where_grad
forward : where (Tensor condition, Tensor x, Tensor y) -> Tensor(out)
args : (Tensor condition, Tensor x, Tensor y, Tensor out_grad)
output : Tensor(x_grad), Tensor(y_grad)
infer_meta :
func : GeneralBinaryGradInferMeta
param : [x, y]
kernel :
func : where_grad
no_need_buffer : x, y
tools/infrt/generate_phi_kernel_dialect.py
浏览文件 @ fcd32950
...@@ -71,9 +71,17 @@ def get_skipped_kernel_list(): ...@@ -71,9 +71,17 @@ def get_skipped_kernel_list():
def get_api_yaml_info(file_path): def get_api_yaml_info(file_path):
f = open(file_path + "/python/paddle/utils/code_gen/api.yaml", "r") apis = []
cont = f.read() with open(file_path + "/python/paddle/utils/code_gen/api.yaml", 'r') as f:
return yaml.load(cont, Loader=yaml.FullLoader) api_list = yaml.load(f, Loader=yaml.FullLoader)
if api_list:
apis.extend(api_list)
with open(file_path + "/python/paddle/utils/code_gen/legacy_api.yaml",
'r') as f:
legacy_api_list = yaml.load(f, Loader=yaml.FullLoader)
if legacy_api_list:
apis.extend(legacy_api_list)
return apis
def generate_kernel_name(op_name, place_str): def generate_kernel_name(op_name, place_str):
......
tools/infrt/get_phi_kernel_function.sh
浏览文件 @ fcd32950
...@@ -78,7 +78,7 @@ done ...@@ -78,7 +78,7 @@ done
#step 2:get simple general inferMeta function wrap info #step 2:get simple general inferMeta function wrap info
temp_path=`mktemp -d` temp_path=`mktemp -d`
python3 ${PADDLE_ROOT}/python/paddle/utils/code_gen/wrapped_infermeta_gen.py \ python3 ${PADDLE_ROOT}/python/paddle/utils/code_gen/wrapped_infermeta_gen.py \
--api_yaml_path ${PADDLE_ROOT}/python/paddle/utils/code_gen/api.yaml \ --api_yaml_path ${PADDLE_ROOT}/python/paddle/utils/code_gen/api.yaml ${PADDLE_ROOT}/python/paddle/utils/code_gen/legacy_api.yaml \
--wrapped_infermeta_header_path ${temp_path}/generate.h \ --wrapped_infermeta_header_path ${temp_path}/generate.h \
--wrapped_infermeta_source_path ${temp_path}/generate.cc --wrapped_infermeta_source_path ${temp_path}/generate.cc
......
tools/infrt/get_phi_kernel_info.py
浏览文件 @ fcd32950
...@@ -21,13 +21,21 @@ from typing import List, Dict, Any ...@@ -21,13 +21,21 @@ from typing import List, Dict, Any
skipped_phi_api_list_file = "/tools/infrt/skipped_phi_api.json" skipped_phi_api_list_file = "/tools/infrt/skipped_phi_api.json"
api_yaml_file = "/python/paddle/utils/code_gen/api.yaml" api_yaml_file = "/python/paddle/utils/code_gen/api.yaml"
legacy_api_yaml_file = "/python/paddle/utils/code_gen/legacy_api.yaml"
def get_skipped_kernel_list(): def get_skipped_kernel_list():
skiped_kernel_list = [] skiped_kernel_list = []
with open(skipped_phi_api_list_file, 'r') as f: with open(skipped_phi_api_list_file, 'r') as f:
skiped_api_list = json.load(f) skiped_api_list = json.load(f)
infer_meta_data = get_api_yaml_info(api_yaml_file) infer_meta_data = []
api_meta_data = get_api_yaml_info(api_yaml_file)
legacy_api_meta_data = get_api_yaml_info(legacy_api_yaml_file)
if api_meta_data:
infer_meta_data.extend(api_meta_data)
if legacy_api_meta_data:
infer_meta_data.extend(legacy_api_meta_data)
for api in infer_meta_data: for api in infer_meta_data:
if "kernel" not in api or "infer_meta" not in api: if "kernel" not in api or "infer_meta" not in api:
continue continue
...@@ -365,7 +373,14 @@ if __name__ == "__main__": ...@@ -365,7 +373,14 @@ if __name__ == "__main__":
args = parse_args() args = parse_args()
skipped_phi_api_list_file = args.paddle_root_path + skipped_phi_api_list_file skipped_phi_api_list_file = args.paddle_root_path + skipped_phi_api_list_file
api_yaml_file = args.paddle_root_path + api_yaml_file api_yaml_file = args.paddle_root_path + api_yaml_file
infer_meta_data = get_api_yaml_info(api_yaml_file) legacy_api_yaml_file = args.paddle_root_path + legacy_api_yaml_file
infer_meta_data = []
api_meta_data = get_api_yaml_info(api_yaml_file)
legacy_api_meta_data = get_api_yaml_info(legacy_api_yaml_file)
if api_meta_data:
infer_meta_data.extend(api_meta_data)
if legacy_api_meta_data:
infer_meta_data.extend(legacy_api_meta_data)
kernel_data = get_kernel_info(args.kernel_info_file) kernel_data = get_kernel_info(args.kernel_info_file)
info_meta_wrap_data = get_infermeta_info(args.infermeta_wrap_file) info_meta_wrap_data = get_infermeta_info(args.infermeta_wrap_file)
attr_data = get_attr_info(args.attr_info_file) attr_data = get_attr_info(args.attr_info_file)
......
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或