From 85489d3938fc22995064b965af946df650c50301 Mon Sep 17 00:00:00 2001 From: zyfncg Date: Wed, 19 Oct 2022 14:15:41 +0800 Subject: [PATCH] Rename name of op and op_args in yaml to align python api (#46343) * rename op in yaml * fix test_layout_autotune * fix layout autotune of transpose --- .../generator/eager_gen.py | 2 +- paddle/phi/api/yaml/backward.yaml | 4 +- paddle/phi/api/yaml/legacy_backward.yaml | 178 +++++++++--------- paddle/phi/api/yaml/legacy_ops.yaml | 110 ++++++----- paddle/phi/api/yaml/op_compat.yaml | 2 +- paddle/phi/api/yaml/ops.yaml | 2 +- python/paddle/fluid/layers/nn.py | 2 +- .../tests/unittests/test_layout_autotune.py | 1 + python/paddle/nn/functional/activation.py | 8 +- python/paddle/tensor/search.py | 2 +- python/paddle/tensor/stat.py | 2 +- python/paddle/vision/ops.py | 2 +- 12 files changed, 157 insertions(+), 158 deletions(-) diff --git a/paddle/fluid/eager/auto_code_generator/generator/eager_gen.py b/paddle/fluid/eager/auto_code_generator/generator/eager_gen.py index 7063524b4d6..cf51fbb9d07 100644 --- a/paddle/fluid/eager/auto_code_generator/generator/eager_gen.py +++ b/paddle/fluid/eager/auto_code_generator/generator/eager_gen.py @@ -1023,7 +1023,7 @@ class DygraphForwardFunctionGenerator(DygraphFunctionGeneratorBase): forward_outputs_position_map.keys()) - len(intermediate_outputs) # for layout autotune attr lightly_sensitive_attr = [ - 'axis', 'axes', 'dim', 'dims', 'start', 'end', 'stop' + 'axis', 'axes', 'dim', 'dims', 'start', 'end', 'stop', 'perm' ] heavily_sensitive_attr = ['data_format', 'data_layout'] layout_autotune_attr = [] diff --git a/paddle/phi/api/yaml/backward.yaml b/paddle/phi/api/yaml/backward.yaml index 7923edac0c8..113f50397cd 100644 --- a/paddle/phi/api/yaml/backward.yaml +++ b/paddle/phi/api/yaml/backward.yaml @@ -217,9 +217,9 @@ no_need_buffer : x - backward_op : trunc_grad - forward : trunc (Tensor x) -> Tensor(out) + forward : trunc (Tensor input) -> Tensor(out) args : (Tensor out_grad) - output : Tensor(x_grad) + output : Tensor(input_grad) infer_meta : func : UnchangedInferMeta param : [out_grad] diff --git a/paddle/phi/api/yaml/legacy_backward.yaml b/paddle/phi/api/yaml/legacy_backward.yaml index aa2b9cf1bcf..a4fea714101 100755 --- a/paddle/phi/api/yaml/legacy_backward.yaml +++ b/paddle/phi/api/yaml/legacy_backward.yaml @@ -105,8 +105,8 @@ use_gpudnn: use_cudnn - backward_op : amax_grad - forward: amax (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) + forward: amax (Tensor x, int64_t[] axis={}, bool keepdim=false) -> Tensor(out) + args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis={}, bool keepdim=false, bool reduce_all=false) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -115,8 +115,8 @@ func : amax_grad - backward_op : amin_grad - forward: amin (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) + forward: amin (Tensor x, int64_t[] axis={}, bool keepdim=false) -> Tensor(out) + args : (Tensor x, Tensor out, Tensor out_grad, int64_t[] axis={}, bool keepdim=false, bool reduce_all=false) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -311,19 +311,19 @@ inplace : (out_grad -> x_grad) - backward_op : broadcast_tensors_grad - forward : broadcast_tensors (Tensor[] x) -> Tensor[](out) - args : (Tensor[] x, Tensor[] out_grad) - output : Tensor[](x_grad) + forward : broadcast_tensors (Tensor[] input) -> Tensor[](out) + args : (Tensor[] input, Tensor[] out_grad) + output : Tensor[](input_grad) infer_meta : func : UnchangedMultiInferMeta - param : [x] + param : [input] kernel : func : broadcast_tensors_grad param : [out_grad] - no_need_buffer : x + no_need_buffer : input - backward_op : cast_grad - forward : cast (Tensor x, DataType out_dtype) -> Tensor(out) + forward : cast (Tensor x, DataType dtype) -> Tensor(out) args : (Tensor x, Tensor out_grad) output : Tensor(x_grad) invoke : cast (out_grad, x.dtype()) @@ -386,14 +386,14 @@ inplace : (out_grad -> x_grad) - backward_op : complex_grad - forward : complex (Tensor x, Tensor y) -> Tensor(out) - args : (Tensor x, Tensor y, Tensor out_grad) - output : Tensor(x_grad), Tensor(y_grad) + forward : complex (Tensor real, Tensor imag) -> Tensor(out) + args : (Tensor real, Tensor imag, Tensor out_grad) + output : Tensor(real_grad), Tensor(imag_grad) infer_meta : func : ComplexGradInferMeta kernel : func : complex_grad - data_type : x + data_type : real - backward_op : concat_double_grad forward : concat_grad (Tensor[] x, Tensor grad_out, Scalar axis) -> Tensor[](grad_x) @@ -663,7 +663,7 @@ skip_transform : out_w, out_w_grad - backward_op : eigh_grad - forward : eigh (Tensor x, str uplo) -> Tensor(out_w), Tensor(out_v) + 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 : @@ -788,7 +788,7 @@ inplace : (out_grad -> x_grad) - backward_op : exponential__grad - forward : exponential_ (Tensor x, float lambda) -> Tensor(out) + forward : exponential_ (Tensor x, float lam) -> Tensor(out) args : (Tensor out_grad) output : Tensor(x_grad) infer_meta : @@ -981,8 +981,8 @@ kernel : func : gumbel_softmax_grad -- backward_op : hard_shrink_grad - forward : hard_shrink (Tensor x, float threshold) -> Tensor(out) +- backward_op : hardshrink_grad + forward : hardshrink (Tensor x, float threshold) -> Tensor(out) args : (Tensor x, Tensor out_grad, float threshold) output : Tensor(x_grad) infer_meta : @@ -992,8 +992,8 @@ func : hard_shrink_grad inplace : (out_grad -> x_grad) -- backward_op : hard_sigmoid_grad - forward : hard_sigmoid (Tensor x, float slope, float offset) -> Tensor(out) +- backward_op : hardsigmoid_grad + forward : hardsigmoid (Tensor x, float slope, float offset) -> Tensor(out) args : (Tensor out, Tensor out_grad, float slope, float offset) output : Tensor(x_grad) infer_meta : @@ -1003,8 +1003,8 @@ func : hard_sigmoid_grad inplace : (out_grad -> x_grad) -- backward_op : hard_swish_grad - forward : hard_swish (Tensor x, float threshold = 6.0, float scale = 6.0, float offset = 3.0) -> Tensor(out) +- backward_op : hardswish_grad + forward : hardswish (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 : @@ -1065,8 +1065,8 @@ no_need_buffer : x - backward_op : index_select_grad - forward : index_select(Tensor x, Tensor index, int dim) -> Tensor(out) - args : (Tensor x, Tensor index, Tensor out_grad, int dim) + forward : index_select(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 @@ -1164,8 +1164,8 @@ optional : scale, bias - backward_op : 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) + forward : leaky_relu_grad (Tensor x, Tensor grad_out, float negative_slope) -> Tensor(grad_x) + args : (Tensor x, Tensor grad_x_grad, float negative_slope) output : Tensor(grad_out_grad) infer_meta : func : UnchangedInferMeta @@ -1175,8 +1175,8 @@ inplace : (grad_x_grad -> grad_out_grad) - backward_op : leaky_relu_grad - forward : leaky_relu (Tensor x, float alpha) -> Tensor(out) - args : (Tensor x, Tensor out_grad, float alpha) + forward : leaky_relu (Tensor x, float negative_slope) -> Tensor(out) + args : (Tensor x, Tensor out_grad, float negative_slope) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -1335,8 +1335,8 @@ func : lu_grad - backward_op : lu_unpack_grad - forward : lu_unpack (Tensor x, Tensor pivots, bool unpack_ludata, bool unpack_pivots) -> Tensor(pmat), Tensor(l), Tensor(u) - args : (Tensor x, Tensor pivots, Tensor l, Tensor u, Tensor pmat, Tensor l_grad, Tensor u_grad, bool unpack_ludata, bool unpack_pivots) + forward : lu_unpack (Tensor x, Tensor y, bool unpack_ludata, bool unpack_pivots) -> Tensor(pmat), Tensor(l), Tensor(u) + args : (Tensor x, Tensor y, Tensor l, Tensor u, Tensor pmat, Tensor l_grad, Tensor u_grad, bool unpack_ludata, bool unpack_pivots) output : Tensor(x_grad) infer_meta : func : LUUnpackGradInferMeta @@ -1411,8 +1411,8 @@ func : matrix_power_grad - backward_op : max_grad - forward: max (Tensor x, IntArray dims={}, bool keep_dim=false) -> Tensor(out) - args : (Tensor x, Tensor out, Tensor out_grad, IntArray dims={}, bool keep_dim=false, bool reduce_all=false) + forward: max (Tensor x, IntArray axis={}, bool keepdim=false) -> Tensor(out) + args : (Tensor x, Tensor out, Tensor out_grad, IntArray axis={}, bool keepdim=false, bool reduce_all=false) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -1469,14 +1469,14 @@ func : mean_all_grad - backward_op : mean_double_grad - forward: mean_grad (Tensor x, Tensor grad_out, IntArray dims={}, bool keep_dim=false, bool reduce_all = false) -> Tensor(grad_x) - args : (Tensor grad_x_grad, IntArray dims={}, bool keep_dim=false) + forward: mean_grad (Tensor x, Tensor grad_out, IntArray axis={}, bool keepdim=false, bool reduce_all = false) -> Tensor(grad_x) + args : (Tensor grad_x_grad, IntArray axis={}, bool keepdim=false) output : Tensor(grad_out_grad) - invoke : mean(grad_x_grad, dims, keep_dim) + invoke : mean(grad_x_grad, axis, keepdim) - backward_op : mean_grad - forward: mean (Tensor x, IntArray dims={}, bool keep_dim=false) -> Tensor(out) - args : (Tensor x, Tensor out_grad, IntArray dims={}, bool keep_dim=false, bool reduce_all=false) + forward: mean (Tensor x, IntArray axis={}, bool keepdim=false) -> Tensor(out) + args : (Tensor x, Tensor out_grad, IntArray axis={}, bool keepdim=false, bool reduce_all=false) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -1496,8 +1496,8 @@ func : meshgrid_grad - backward_op : min_grad - forward: min (Tensor x, IntArray dims={}, bool keep_dim=false) -> Tensor(out) - args : (Tensor x, Tensor out, Tensor out_grad, IntArray dims={}, bool keep_dim=false, bool reduce_all=false) + forward: min (Tensor x, IntArray axis={}, bool keepdim=false) -> Tensor(out) + args : (Tensor x, Tensor out, Tensor out_grad, IntArray axis={}, bool keepdim=false, bool reduce_all=false) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -1546,15 +1546,15 @@ func : multi_dot_grad - backward_op : multiplex_grad - forward : multiplex (Tensor[] ins, Tensor ids) -> Tensor(out) - args : (Tensor[] ins, Tensor ids, Tensor out_grad) - output : Tensor[](ins_grad){ins.size()} + forward : multiplex (Tensor[] inputs, Tensor index) -> Tensor(out) + args : (Tensor[] inputs, Tensor index, Tensor out_grad) + output : Tensor[](inputs_grad){inputs.size()} infer_meta : func : MultiplexGradInferMeta - param : [ids, out_grad] + param : [index, out_grad] kernel : func : multiplex_grad - param : [ids, out_grad] + param : [index, out_grad] - backward_op : multiply_double_grad forward : multiply_grad (Tensor x, Tensor y, Tensor grad_out, int axis = -1) -> Tensor(grad_x), Tensor(grad_y) @@ -1734,8 +1734,8 @@ use_gpudnn : use_gpudnn - backward_op : pow_grad - forward : pow(Tensor x, Scalar s) -> Tensor(out) - args : (Tensor x, Tensor out_grad, Scalar s=-1) + forward : pow(Tensor x, Scalar y) -> Tensor(out) + args : (Tensor x, Tensor out_grad, Scalar y=-1) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -1768,12 +1768,12 @@ # output is optional - backward_op : 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) + forward : put_along_axis (Tensor arr, Tensor index, Tensor value, int axis, str reduce) -> Tensor(out) + args : (Tensor arr, Tensor index, Tensor out_grad, int axis, str reduce) + output : Tensor(arr_grad), Tensor(value_grad) infer_meta : func : GeneralBinaryGradInferMeta - param : [x, index] + param : [arr, index] kernel : func : put_along_axis_grad @@ -1859,8 +1859,8 @@ func : renorm_grad - backward_op : repeat_interleave_grad - forward : repeat_interleave(Tensor x, int repeats, int dim) -> Tensor(out) - args : (Tensor x, Tensor out_grad, int repeats, int dim) + forward : repeat_interleave(Tensor x, int repeats, int axis) -> Tensor(out) + args : (Tensor x, Tensor out_grad, int repeats, int axis) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -1869,8 +1869,8 @@ func : repeat_interleave_grad - backward_op : repeat_interleave_with_tensor_index_grad - forward : repeat_interleave_with_tensor_index(Tensor x, Tensor repeats, int dim) -> Tensor(out) - args : (Tensor x, Tensor repeats, Tensor out_grad, int dim) + forward : repeat_interleave_with_tensor_index(Tensor x, Tensor repeats, int axis) -> Tensor(out) + args : (Tensor x, Tensor repeats, Tensor out_grad, int axis) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -2169,17 +2169,6 @@ kernel : func : slogdeterminant_grad -- backward_op : 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_op : softmax_grad forward : softmax (Tensor x, int axis) -> Tensor(out) args : (Tensor out, Tensor out_grad, int axis) @@ -2202,6 +2191,17 @@ func : softplus_grad inplace : (out_grad -> x_grad) +- backward_op : softshrink_grad + forward : softshrink (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 : soft_shrink_grad + inplace : (out_grad -> x_grad) + - backward_op : softsign_grad forward : softsign (Tensor x) -> Tensor(out) args : (Tensor x, Tensor out_grad) @@ -2293,14 +2293,14 @@ func : squared_l2_norm_grad - backward_op : squeeze_double_grad - forward : squeeze_grad(Tensor xshape, Tensor grad_out, IntArray axes) -> Tensor(grad_x) - args : (Tensor grad_x_grad, IntArray axes) + forward : squeeze_grad(Tensor xshape, Tensor grad_out, IntArray axis) -> Tensor(grad_x) + args : (Tensor grad_x_grad, IntArray axis) output : Tensor(grad_out_grad) - invoke: squeeze(grad_x_grad, axes) + invoke: squeeze(grad_x_grad, axis) - backward_op : squeeze_grad - forward : squeeze(Tensor x, IntArray axes) -> Tensor(out), Tensor(xshape) - args : (Tensor xshape, Tensor out_grad, IntArray axes) + forward : squeeze(Tensor x, IntArray axis) -> Tensor(out), Tensor(xshape) + args : (Tensor xshape, Tensor out_grad, IntArray axis) output : Tensor(x_grad) infer_meta : func : KernelWithXShapeInferMeta @@ -2360,14 +2360,14 @@ inplace : (out_grad -> x_grad) - backward_op : sum_double_grad - forward : sum_grad (Tensor x, Tensor grad_out, IntArray dims, bool keep_dim, bool reduce_all=false) -> Tensor(grad_x) - args : (Tensor grad_x_grad, IntArray dims={}, bool keep_dim=false) + forward : sum_grad (Tensor x, Tensor grad_out, IntArray axis, bool keepdim, bool reduce_all=false) -> Tensor(grad_x) + args : (Tensor grad_x_grad, IntArray axis={}, bool keepdim=false) output : Tensor(grad_out_grad) - invoke : sum(grad_x_grad, dims, grad_x_grad.dtype(), keep_dim) + invoke : sum(grad_x_grad, axis, grad_x_grad.dtype(), keepdim) - backward_op : sum_grad - forward : sum (Tensor x, IntArray dims={}, DataType out_dtype=DataType::UNDEFINED, bool keep_dim=false) -> Tensor(out) - args : (Tensor x, Tensor out_grad, IntArray dims, bool keep_dim, bool reduce_all=false) + forward : sum (Tensor x, IntArray axis={}, DataType dtype=DataType::UNDEFINED, bool keepdim=false) -> Tensor(out) + args : (Tensor x, Tensor out_grad, IntArray axis, bool keepdim, bool reduce_all=false) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -2378,8 +2378,8 @@ backward : sum_double_grad - backward_op : svd_grad - forward : svd (Tensor x, bool full) -> Tensor(u), Tensor(s), Tensor(vh) - args : (Tensor x, Tensor u, Tensor vh, Tensor s, Tensor u_grad, Tensor vh_grad, Tensor s_grad, bool full) + forward : svd (Tensor x, bool full_matrices) -> Tensor(u), Tensor(s), Tensor(vh) + args : (Tensor x, Tensor u, Tensor vh, Tensor s, Tensor u_grad, Tensor vh_grad, Tensor s_grad, bool full_matrices) output : Tensor(x_grad) infer_meta : func : UnchangedInferMeta @@ -2412,12 +2412,12 @@ optional : reserve_space - backward_op : 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) + forward : take_along_axis (Tensor arr, Tensor indices, int axis) -> Tensor(out) + args : (Tensor arr, Tensor indices, Tensor out_grad, int axis) + output : Tensor(arr_grad) infer_meta : func : UnchangedInferMeta - param : [x] + param : [arr] kernel : func : take_along_axis_grad @@ -2517,8 +2517,8 @@ no_need_buffer : x backward : tile_double_grad -- backward_op : top_k_grad - forward : top_k (Tensor x, Scalar k, int axis = -1, bool largest = true, bool sorted = true) -> Tensor(out), Tensor(indices) +- backward_op : topk_grad + forward : topk (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 : @@ -2528,18 +2528,18 @@ func : top_k_grad - backward_op : transpose_double_grad - forward : transpose_grad (Tensor grad_out, int[] axis) -> Tensor(grad_x) - args : (Tensor grad_x_grad, int[] axis) + forward : transpose_grad (Tensor grad_out, int[] perm) -> Tensor(grad_x) + args : (Tensor grad_x_grad, int[] perm) output : Tensor(grad_out_grad) - invoke : transpose(grad_x_grad, axis) + invoke : transpose(grad_x_grad, perm) - backward_op : transpose_grad - forward : transpose (Tensor x, int[] axis) -> Tensor(out) - args : (Tensor out_grad, int[] axis) + forward : transpose (Tensor x, int[] perm) -> Tensor(out) + args : (Tensor out_grad, int[] perm) output : Tensor(x_grad) infer_meta : func : TransposeGradInferMeta - param : [out_grad, axis] + param : [out_grad, perm] kernel : func : transpose_grad backward : transpose_double_grad diff --git a/paddle/phi/api/yaml/legacy_ops.yaml b/paddle/phi/api/yaml/legacy_ops.yaml index efd585db488..0d37d3e76f6 100755 --- a/paddle/phi/api/yaml/legacy_ops.yaml +++ b/paddle/phi/api/yaml/legacy_ops.yaml @@ -100,9 +100,9 @@ backward : add_grad - op : add_n - args : (Tensor[] x) + args : (Tensor[] inputs) output : Tensor - invoke : add_n_impl(x) + invoke : add_n_impl(inputs) backward : add_n_grad - op : addmm @@ -128,7 +128,7 @@ backward : affine_grid_grad - op : all - args : (Tensor x, int64_t[] dims={}, bool keep_dim=false) + args : (Tensor x, int64_t[] axis={}, bool keepdim=false) output : Tensor(out) infer_meta : func : ReduceInferMeta @@ -145,7 +145,7 @@ func : allclose - op : amax - args : (Tensor x, int64_t[] dims={}, bool keep_dim=false) + args : (Tensor x, int64_t[] axis={}, bool keepdim=false) output : Tensor(out) infer_meta : func : ReduceInferMeta @@ -154,7 +154,7 @@ backward : amax_grad - op : amin - args : (Tensor x, int64_t[] dims={}, bool keep_dim=false) + args : (Tensor x, int64_t[] axis={}, bool keepdim=false) output : Tensor(out) infer_meta : func : ReduceInferMeta @@ -172,7 +172,7 @@ backward : angle_grad - op : any - args : (Tensor x, int64_t[] dims={}, bool keep_dim=false) + args : (Tensor x, int64_t[] axis={}, bool keepdim=false) output : Tensor(out) infer_meta : func : ReduceInferMeta @@ -438,13 +438,13 @@ backward : brelu_grad - op : cast - args : (Tensor x, DataType out_dtype) + args : (Tensor x, DataType dtype) output : Tensor infer_meta : func : CastInferMeta kernel : func : cast - param : [x, out_dtype] + param : [x, dtype] data_type : x backward : cast_grad @@ -517,7 +517,7 @@ data_type : dtype - op : complex - args : (Tensor x, Tensor y) + args : (Tensor real, Tensor imag) output : Tensor infer_meta : func : ComplexInferMeta @@ -700,7 +700,7 @@ backward : det_grad - op : diag_embed - args : (Tensor x, int offset, int dim1, int dim2) + args : (Tensor input, int offset, int dim1, int dim2) output : Tensor(out) infer_meta : func : DiagEmbedInferMeta @@ -748,7 +748,7 @@ optional : hypslength, refslength - op : eigh - args : (Tensor x, str uplo) + args : (Tensor x, str UPLO) output : Tensor(out_w), Tensor(out_v) infer_meta : func : EighInferMeta @@ -896,7 +896,7 @@ backward : expm1_grad - op : exponential_ - args : (Tensor x, float lambda) + args : (Tensor x, float lam) output : Tensor(out) infer_meta : func : UnchangedInferMeta @@ -1119,7 +1119,7 @@ func : gelu backward : gelu_grad -- op : generate_proposals_v2 +- op : generate_proposals args : (Tensor scores, Tensor bbox_deltas, Tensor im_shape, Tensor anchors, Tensor variances, int pre_nms_top_n, int post_nms_top_n, float nms_thresh, float min_size, float eta, bool pixel_offset=true) output : Tensor(rpn_rois), Tensor(rpn_roi_probs), Tensor(rpn_rois_num) infer_meta : @@ -1196,7 +1196,7 @@ func : gumbel_softmax backward : gumbel_softmax_grad -- op : hard_shrink +- op : hardshrink args : (Tensor x, float threshold) output : Tensor infer_meta : @@ -1204,9 +1204,9 @@ param : [x] kernel : func : hard_shrink - backward : hard_shrink_grad + backward : hardshrink_grad -- op : hard_sigmoid +- op : hardsigmoid args : (Tensor x, float slope, float offset) output : Tensor infer_meta : @@ -1214,9 +1214,9 @@ param : [x] kernel : func : hard_sigmoid - backward : hard_sigmoid_grad + backward : hardsigmoid_grad -- op : hard_swish +- op : hardswish args : (Tensor x, float threshold = 6.0, float scale = 6.0, float offset = 3.0) output : Tensor infer_meta : @@ -1224,7 +1224,7 @@ param : [x] kernel : func : hard_swish - backward : hard_swish_grad + backward : hardswish_grad - op : hierarchical_sigmoid args : (Tensor x, Tensor w, Tensor label, Tensor path, Tensor code, Tensor bias, int num_classes, bool remote_prefetch, int trainer_id, int64_t[] height_sections, str[] epmap, str[] table_names, bool is_sparse) @@ -1238,7 +1238,7 @@ backward : hierarchical_sigmoid_grad - op : histogram - args : (Tensor x, int64_t bins, int min, int max) + args : (Tensor input, int64_t bins, int min, int max) output : Tensor(out) infer_meta : func : HistogramInferMeta @@ -1294,7 +1294,7 @@ backward : index_sample_grad - op : index_select - args : (Tensor x, Tensor index, int dim) + args : (Tensor x, Tensor index, int axis) output : Tensor(out) infer_meta : func : IndexSelectInferMeta @@ -1432,7 +1432,7 @@ optional : scale, bias - op : leaky_relu - args : (Tensor x, float alpha) + args : (Tensor x, float negative_slope) output : Tensor infer_meta : func : UnchangedInferMeta @@ -1632,7 +1632,7 @@ backward : lu_grad - op : lu_unpack - args : (Tensor x, Tensor pivots, bool unpack_ludata, bool unpack_pivots) + args : (Tensor x, Tensor y, bool unpack_ludata, bool unpack_pivots) output : Tensor(pmat), Tensor(l), Tensor(u) infer_meta : func : LUUnpackInferMeta @@ -1706,7 +1706,7 @@ func : matrix_rank_tol - op : max - args : (Tensor x, IntArray dims={}, bool keep_dim=false) + args : (Tensor x, IntArray axis={}, bool keepdim=false) output : Tensor(out) infer_meta : func : ReduceIntArrayAxisInferMeta @@ -1751,7 +1751,7 @@ backward : maxout_grad - op : mean - args : (Tensor x, IntArray dims={}, bool keep_dim=false) + args : (Tensor x, IntArray axis={}, bool keepdim=false) output : Tensor(out) infer_meta : func : ReduceIntArrayAxisInferMeta @@ -1808,7 +1808,7 @@ backward : meshgrid_grad - op : min - args : (Tensor x, IntArray dims={}, bool keep_dim=false) + args : (Tensor x, IntArray axis={}, bool keepdim=false) output : Tensor(out) infer_meta : func : ReduceIntArrayAxisInferMeta @@ -1882,13 +1882,13 @@ func : multinomial - op : multiplex - args : (Tensor[] ins, Tensor ids) + args : (Tensor[] inputs, Tensor index) output : Tensor infer_meta : func : MultiplexInferMeta kernel : func : multiplex - data_type : ins + data_type : inputs backward : multiplex_grad - op : multiply @@ -2028,7 +2028,7 @@ backward : pool3d_grad - op : pow - args : (Tensor x, Scalar s) + args : (Tensor x, Scalar y) output : Tensor(out) infer_meta : func : UnchangedInferMeta @@ -2066,15 +2066,15 @@ backward : psroi_pool_grad - op : put_along_axis - args : (Tensor x, Tensor index, Tensor value, int axis, str reduce) + args : (Tensor arr, Tensor index, Tensor value, int axis, str reduce) output : Tensor(out) infer_meta : func : UnchangedInferMeta - param : [x] + param : [arr] kernel : func : put_along_axis - data_type : x - inplace : (x -> out) + data_type : arr + inplace : (arr -> out) backward : put_along_axis_grad - op : qr @@ -2178,21 +2178,19 @@ backward : renorm_grad - op : repeat_interleave - args : (Tensor x, int repeats, int dim) + args : (Tensor x, int repeats, int axis) output : Tensor(out) infer_meta : func : RepeatInterleaveInferMeta - param : [x,repeats, dim] kernel : func : repeat_interleave backward: repeat_interleave_grad - op : repeat_interleave_with_tensor_index - args : (Tensor x, Tensor repeats, int dim) + args : (Tensor x, Tensor repeats, int axis) output : Tensor(out) infer_meta : func : RepeatInterleaveWithTensorIndexInferMeta - param : [x,repeats, dim] kernel : func : repeat_interleave_with_tensor_index data_type : x @@ -2316,7 +2314,7 @@ backward : scatter_nd_add_grad - op : searchsorted - args : (Tensor sorted_sequence, Tensor value, bool out_int32, bool right) + args : (Tensor sorted_sequence, Tensor values, bool out_int32, bool right) output : Tensor(out) infer_meta : func : SearchsortedInferMeta @@ -2371,7 +2369,7 @@ skip_transform : input - op : shard_index - args : (Tensor in, int index_num, int nshards, int shard_id, int ignore_value) + args : (Tensor input, int index_num, int nshards, int shard_id, int ignore_value) output : Tensor(out) infer_meta : func : ShardIndexInferMeta @@ -2432,7 +2430,7 @@ func : sinh backward : sinh_grad -- op : size +- op : numel args : (Tensor x) output : Tensor(size) infer_meta : @@ -2460,15 +2458,15 @@ func : slogdeterminant backward : slogdet_grad -- op : soft_shrink - args : (Tensor x, float lambda) +- op : softshrink + args : (Tensor x, float threshold) output : Tensor infer_meta : func : UnchangedInferMeta param : [x] kernel : func : soft_shrink - backward : soft_shrink_grad + backward : softshrink_grad - op : softmax args : (Tensor x, int axis) @@ -2558,7 +2556,7 @@ backward : squared_l2_norm_grad - op : squeeze - args : (Tensor x, IntArray axes) + args : (Tensor x, IntArray axis) output : Tensor(out), Tensor(xshape) infer_meta : func : SqueezeWithXShapeInferMeta @@ -2598,7 +2596,7 @@ backward : subtract_grad - op : sum - args : (Tensor x, IntArray dims={}, DataType out_dtype=DataType::UNDEFINED, bool keep_dim=false) + args : (Tensor x, IntArray axis={}, DataType dtype=DataType::UNDEFINED, bool keepdim=false) output : Tensor(out) infer_meta : func : SumInferMeta @@ -2608,7 +2606,7 @@ backward : sum_grad - op : svd - args : (Tensor x, bool full_metrices) + args : (Tensor x, bool full_matrices) output : Tensor(u), Tensor(s), Tensor(vh) infer_meta : func : SvdInferMeta @@ -2639,14 +2637,14 @@ inplace : (mean -> mean_out), (variance -> variance_out) - op : take_along_axis - args : (Tensor x, Tensor index, int axis) + args : (Tensor arr, Tensor indices, int axis) output : Tensor infer_meta : func : UnchangedInferMeta - param : [index] + param : [indices] kernel : func : take_along_axis - data_type : x + data_type : arr backward : take_along_axis_grad - op : tan @@ -2705,17 +2703,17 @@ func : tile backward : tile_grad -- op : top_k +- op : topk 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 + backward : topk_grad - op : transpose - args : (Tensor x, int[] axis) + args : (Tensor x, int[] perm) output : Tensor infer_meta : func : TransposeInferMeta @@ -2871,13 +2869,13 @@ backward : unstack_grad - op : viterbi_decode - args : (Tensor input, Tensor transition, Tensor length, bool include_bos_eos_tag) + args : (Tensor potentials, Tensor transition_params, Tensor lengths, bool include_bos_eos_tag) output : Tensor(scores), Tensor(path) infer_meta : func : ViterbiDecodeInferMeta kernel : func : viterbi_decode - data_type : input + data_type : potentials - op : warpctc args : (Tensor logits, Tensor label, Tensor logits_length, Tensor labels_length, int blank, bool norm_by_times) @@ -2939,8 +2937,8 @@ invoke : full_like(x, 0, dtype, place) - op: broadcast_tensors - args: (Tensor[] x) - output: Tensor[]{x.size()} + args: (Tensor[] input) + output: Tensor[]{input.size()} infer_meta: func: BroadcastTensorsInferMeta kernel: diff --git a/paddle/phi/api/yaml/op_compat.yaml b/paddle/phi/api/yaml/op_compat.yaml index bcb3563a040..29e88420262 100644 --- a/paddle/phi/api/yaml/op_compat.yaml +++ b/paddle/phi/api/yaml/op_compat.yaml @@ -774,7 +774,7 @@ - op : trunc inputs : - x : X + input : X outputs : out : Out diff --git a/paddle/phi/api/yaml/ops.yaml b/paddle/phi/api/yaml/ops.yaml index 10e617bd912..ec16844fc49 100644 --- a/paddle/phi/api/yaml/ops.yaml +++ b/paddle/phi/api/yaml/ops.yaml @@ -192,7 +192,7 @@ backward : trace_grad - op : trunc - args : (Tensor x) + args : (Tensor input) output : Tensor infer_meta : func : UnchangedInferMeta diff --git a/python/paddle/fluid/layers/nn.py b/python/paddle/fluid/layers/nn.py index b7814bc19c4..93021fe91bf 100755 --- a/python/paddle/fluid/layers/nn.py +++ b/python/paddle/fluid/layers/nn.py @@ -12009,7 +12009,7 @@ def size(input): """ if in_dygraph_mode(): - return _C_ops.size(input) + return _C_ops.numel(input) if _in_legacy_dygraph(): return _legacy_C_ops.size(input) diff --git a/python/paddle/fluid/tests/unittests/test_layout_autotune.py b/python/paddle/fluid/tests/unittests/test_layout_autotune.py index e004850587f..b7af0464ba9 100644 --- a/python/paddle/fluid/tests/unittests/test_layout_autotune.py +++ b/python/paddle/fluid/tests/unittests/test_layout_autotune.py @@ -51,6 +51,7 @@ class LayoutAutoTune(unittest.TestCase): self.assertEqual(paddle.fluid.core.use_layout_autotune(), True) paddle.fluid.core.disable_layout_autotune() self.assertEqual(paddle.fluid.core.use_layout_autotune(), False) + self.use_autoune() def setUp(self): self.use_autoune() diff --git a/python/paddle/nn/functional/activation.py b/python/paddle/nn/functional/activation.py index bcde1f06496..929fd243715 100644 --- a/python/paddle/nn/functional/activation.py +++ b/python/paddle/nn/functional/activation.py @@ -228,7 +228,7 @@ def hardshrink(x, threshold=0.5, name=None): """ if in_dygraph_mode(): - return _C_ops.hard_shrink(x, threshold) + return _C_ops.hardshrink(x, threshold) if _in_legacy_dygraph(): return _legacy_C_ops.hard_shrink(x, 'threshold', threshold) @@ -336,7 +336,7 @@ def hardsigmoid(x, slope=0.1666667, offset=0.5, name=None): """ if in_dygraph_mode(): - return _C_ops.hard_sigmoid(x, slope, offset) + return _C_ops.hardsigmoid(x, slope, offset) if _in_legacy_dygraph(): return _legacy_C_ops.hard_sigmoid(x, 'slope', slope, 'offset', offset) @@ -393,7 +393,7 @@ def hardswish(x, name=None): if _in_legacy_dygraph(): return _legacy_C_ops.hard_swish(x) if in_dygraph_mode(): - return _C_ops.hard_swish(x, 6, 6, 3) + return _C_ops.hardswish(x, 6, 6, 3) check_variable_and_dtype(x, 'x', ['float16', 'float32', 'float64'], 'hardswish') @@ -1248,7 +1248,7 @@ def softshrink(x, threshold=0.5, name=None): threshold)) if in_dygraph_mode(): - return _C_ops.soft_shrink(x, threshold) + return _C_ops.softshrink(x, threshold) if _in_legacy_dygraph(): return _legacy_C_ops.softshrink(x, 'lambda', threshold) diff --git a/python/paddle/tensor/search.py b/python/paddle/tensor/search.py index 0584ae8dcdd..ffb6573c9eb 100644 --- a/python/paddle/tensor/search.py +++ b/python/paddle/tensor/search.py @@ -865,7 +865,7 @@ def topk(x, k, axis=None, largest=True, sorted=True, name=None): if in_dygraph_mode(): if axis == None: axis = -1 - out, indices = _C_ops.top_k(x, k, axis, largest, sorted) + out, indices = _C_ops.topk(x, k, axis, largest, sorted) return out, indices if _non_static_mode(): diff --git a/python/paddle/tensor/stat.py b/python/paddle/tensor/stat.py index a22a2649719..bbe98c6ab7e 100644 --- a/python/paddle/tensor/stat.py +++ b/python/paddle/tensor/stat.py @@ -244,7 +244,7 @@ def numel(x, name=None): """ if in_dygraph_mode(): - return _C_ops.size(x) + return _C_ops.numel(x) elif _in_legacy_dygraph(): return _legacy_C_ops.size(x) diff --git a/python/paddle/vision/ops.py b/python/paddle/vision/ops.py index 112d233cc49..b66db2aa737 100755 --- a/python/paddle/vision/ops.py +++ b/python/paddle/vision/ops.py @@ -1736,7 +1736,7 @@ def generate_proposals(scores, assert return_rois_num, "return_rois_num should be True in dygraph mode." attrs = (pre_nms_top_n, post_nms_top_n, nms_thresh, min_size, eta, pixel_offset) - rpn_rois, rpn_roi_probs, rpn_rois_num = _C_ops.generate_proposals_v2( + rpn_rois, rpn_roi_probs, rpn_rois_num = _C_ops.generate_proposals( scores, bbox_deltas, img_size, anchors, variances, *attrs) return rpn_rois, rpn_roi_probs, rpn_rois_num -- GitLab