Movatterモバイル変換


[0]ホーム

URL:


Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Sign up
Appearance settings

PyTorch Ops to oneDNN Functions Mapping

Jing Xu edited this pageNov 16, 2023 ·1 revision

PyTorch uses ops that are registered to corresponding Math Kernel Library (MKL) functions in oneDNN. The available implementations are defined in this YAML filenative_functions.yaml in the aten library of PyTorch. By doing a search for the keyword “mkldnn”, all the mappings can be found.

This is summarized in the following table:

PyTorchOp oneDNN Function
add.Tensormkldnn_add
add_.Tensormkldnn_add_
add.outmkldnn_add_out
copy_copy_mkldnn_
empty.memory_formatempty_mkldnn
mkldnn_linearmkldnn_linear
mkldnn_linear_backward_inputmkldnn_linear_backward_input
mkldnn_linear_backward_weightsmkldnn_linear_backward_weights
mkldnn_linear_backwardmkldnn_linear_backward
mkldnn_max_pool2dmkldnn_max_pool2d
mkldnn_max_pool2d_backwardmkldnn_max_pool2d_backward
mkldnn_max_pool3dmkldnn_max_pool3d
mkldnn_max_pool3d_backwardmkldnn_max_pool3d_backward
mkldnn_convolutionmkldnn_convolution
mkldnn_rnn_layermkldnn_rnn_layer
mkldnn_rnn_layer_backwardmkldnn_rnn_layer_backward
mul.Tensormkldnn_mul
mul_.Tensormkldnn_mul_
mul.outmkldnn_mul_out
native_batch_normmkldnn_batch_norm
_native_batch_norm_legit_mkldnn_batch_norm_legit
_native_batch_norm_legit.no_stats_mkldnn_batch_norm_legit_no_stats
native_batch_norm_backwardmkldnn_batch_norm_backward
_mkldnn_reshapemkldnn_reshape
relumkldnn_relu
relu_mkldnn_relu_
_prelu_kernelmkldnn_prelu_backward
gelumkldnn_gelu
gelu_backwardmkldnn_gelu_backward
sigmoidmkldnn_sigmoid
sigmoid_mkldnn_sigmoid_
_softmaxmkldnn_softmax
tanhmkldnn_tanh
tanh_mkldnn_tanh_
threshold_backwardmkldnn_relu_backward
_mkldnn_transposemkldnn_transpose
_mkldnn_transpose_mkldnn_transpose_
clonemkldnn_clone
zero_mkldnn_zero_
_to_densemkldnn_to_dense
to_mkldnndense_to_mkldnn
mkldnn_reorder_conv2d_weightmkldnn_reorder_conv2d_weight
mkldnn_reorder_conv3d_weightmkldnn_reorder_conv3d_weight
viewmkldnn_view
adaptive_avg_pool2d.outmkldnn_adaptive_avg_pool2d_out_stub
mkldnn_adaptive_avg_pool2dmkldnn_adaptive_avg_pool2d
mkldnn_adaptive_avg_pool2d.outmkldnn_adaptive_avg_pool2d_out
mkldnn_adaptive_avg_pool2d_backwardmkldnn_adaptive_avg_pool2d_backward
avg_pool2d.outmkldnn_avg_pool2d_out
avg_pool2dmkldnn_avg_pool2d
avg_pool2d_backward.grad_inputmkldnn_avg_pool2d_backward_out
avg_pool2d_backwardmkldnn_avg_pool2d_backward
avg_pool3d.outmkldnn_avg_pool3d_out
avg_pool3dmkldnn_avg_pool3d
avg_pool3d_backward.grad_inputmkldnn_avg_pool3d_backward_out
avg_pool3d_backwardmkldnn_avg_pool3d_backward

I would love to contribute to PyTorch!

Clone this wiki locally


[8]ページ先頭

©2009-2025 Movatter.jp