mindspore-Ecosystem
/
mindspore
21
0
Fork 27
Code Issues Pull Requests 6 Projects Releases Wiki Activity

add op for ocr

This commit is contained in:
xiao_yao1994 2022-08-29 20:54:43 +08:00
parent 159142a1f4
commit 9d5f70d97a
31 changed files with 476 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.jenkins/check/config/filter_pylint.txt
View File

@ -19,6 +19,7 @@
"mindspore/mindspore/python/mindspore/ops/operations" "super-init-not-called"
"mindspore/mindspore/python/mindspore/ops/operations/_quant_ops.py" "unused-import"
"mindspore/mindspore/python/mindspore/ops/operations/nn_ops.py" "redefined-builtin"
"mindspore/mindspore/python/mindspore/ops/operations/_inner_ops.py" "dangerous-default-value"
"mindspore/mindspore/python/mindspore/ops/operations/_thor_ops.py" "dangerous-default-value"
"mindspore/mindspore/python/mindspore/ops/operations/_thor_ops.py" "redefined-outer-name"
"mindspore/mindspore/python/mindspore/ops/operations/_thor_ops.py" "unused-import"

17
mindspore/ccsrc/transform/graph_ir/op_adapter_map.h
View File

@ -101,6 +101,7 @@ constexpr const char kNameElu[] = "Elu";
constexpr const char kNameEluGrad[] = "EluGrad";
constexpr const char kNameTensorScatterUpdate[] = "TensorScatterUpdate";
constexpr const char kNameTensorScatterElements[] = "TensorScatterElements";
constexpr const char kNameScatterElements[] = "ScatterElements";
constexpr const char kNameNonZero[] = "NonZero";
constexpr const char kNameNonZeroWithValue[] = "NonZeroWithValue";
constexpr const char kNameNonZeroWithValueShape[] = "NonZeroWithValueShape";
@ -356,7 +357,6 @@ constexpr const char kNameWhile[] = "While";
constexpr const char kNameKMeansCentroids[] = "KMeansCentroids";
constexpr const char kNameIsNan[] = "IsNan";
constexpr const char kNameKLDiv[] = "KLDivLoss";
constexpr const char kNameStringLength[] = "StringLength";
constexpr const char kNameGetShape[] = "GetShape";
constexpr const char kNameKlDivLossGrad[] = "KLDivLossGrad";
constexpr const char kNameRandomStandardNormal[] = "RandomStandardNormal";
@ -364,6 +364,21 @@ constexpr const char kNameUnsortedSegmentSum[] = "UnsortedSegmentSum";
constexpr const char kNameSpaceToBatchTF[] = "SpaceToBatchTF";
constexpr const char kNameBatchToSpaceTF[] = "BatchToSpaceTF";
constexpr const char kNameMaskedSelect[] = "MaskedSelect";
constexpr const char kNamePartitionedCall[] = "PartitionedCall";
constexpr const char kNameRangeV2[] = "RangeV2";
constexpr const char kNameOCRDetectionPreHandle[] = "OCRDetectionPreHandle";
constexpr const char kNameOCRFindContours[] = "OCRFindContours";
constexpr const char kNameBatchDilatePolys[] = "BatchDilatePolys";
constexpr const char kNameResizeAndClipPolys[] = "ResizeAndClipPolys";
constexpr const char kNameOCRDetectionPostHandle[] = "OCRDetectionPostHandle";
constexpr const char kNameOCRIdentifyPreHandle[] = "OCRIdentifyPreHandle";
constexpr const char kNameBatchEnqueue[] = "BatchEnqueue";
constexpr const char kNameDequeue[] = "Dequeue";
constexpr const char kNameOCRRecognitionPreHandle[] = "OCRRecognitionPreHandle";
constexpr const char kNameStringUpper[] = "StringUpper";
constexpr const char kNameStringLength[] = "StringLength";
constexpr const char kNameDecodeImage[] = "DecodeImage";
constexpr const char kNameDecodeBase64[] = "DecodeBase64";
class OpAdapterDesc;

7
mindspore/ccsrc/transform/graph_ir/op_adapter_util.cc
View File

@ -189,6 +189,13 @@ GeTensor ConvertAnyUtil(const ValuePtr &value, const AnyTraits<AnyValue>) {
auto v = GetValue<int32_t>(value);
desc.SetRealDimCnt(0);
return GeTensor(desc, reinterpret_cast<uint8_t *>(&v), sizeof(int32_t));
} else if (value->isa<UInt32Imm>()) {
// convert scalar UInt to GeTensor
MS_LOG(INFO) << "Convert scalar to tensor with data type = UInt32";
GeTensorDesc desc(GeShape(), ::ge::FORMAT_NCHW, ::ge::DT_UINT32);
auto v = GetValue<uint32_t>(value);
desc.SetRealDimCnt(0);
return GeTensor(desc, reinterpret_cast<uint8_t *>(&v), sizeof(uint32_t));
} else if (value->isa<Int64Imm>()) {
// convert scalar Int64 to GeTensor
MS_LOG(INFO) << "convert scalar to tensor with data type = Int64";

6
mindspore/ccsrc/transform/graph_ir/op_declare/array_ops_declare.cc
View File

@ -85,6 +85,12 @@ ATTR_MAP(MirrorPadGrad) = {{"mode", ATTR_DESC(mode, AnyTraits<std::string>())}};
OUTPUT_MAP(MirrorPadGrad) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(MirrorPadGrad, kNameMirrorPadGrad, ADPT_DESC(MirrorPadGrad))
// Expand
INPUT_MAP(Expand) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(shape)}};
ATTR_MAP(Expand) = EMPTY_ATTR_MAP;
OUTPUT_MAP(Expand) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(Expand, "Expand", ADPT_DESC(Expand))
// ExpandDims
INPUT_MAP(ExpandDims) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(axis)}};
ATTR_MAP(ExpandDims) = EMPTY_ATTR_MAP;

3
mindspore/ccsrc/transform/graph_ir/op_declare/array_ops_declare.h
View File

@ -42,6 +42,9 @@ DECLARE_OP_USE_OUTPUT(MirrorPad)
DECLARE_OP_ADAPTER(MirrorPadGrad)
DECLARE_OP_USE_OUTPUT(MirrorPadGrad)
DECLARE_OP_ADAPTER(Expand)
DECLARE_OP_USE_OUTPUT(Expand)
DECLARE_OP_ADAPTER(ExpandDims)
DECLARE_OP_USE_OUTPUT(ExpandDims)

5
mindspore/ccsrc/transform/graph_ir/op_declare/elewise_calculation_ops_declare.cc
View File

@ -20,6 +20,11 @@
#include <string>
namespace mindspore::transform {
INPUT_MAP(ClipByValue) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(clip_value_min)}, {3, INPUT_DESC(clip_value_max)}};
ATTR_MAP(ClipByValue) = EMPTY_ATTR_MAP;
OUTPUT_MAP(ClipByValue) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(ClipByValue, "Clip", ADPT_DESC(ClipByValue))
// Assign
INPUT_MAP(Assign) = {{1, INPUT_DESC(ref)}, {2, INPUT_DESC(value)}};
ATTR_MAP(Assign) = EMPTY_ATTR_MAP;

3
mindspore/ccsrc/transform/graph_ir/op_declare/elewise_calculation_ops_declare.h
View File

@ -22,6 +22,9 @@
#include "ops/elewise_calculation_ops.h"
namespace mindspore::transform {
DECLARE_OP_ADAPTER(ClipByValue)
DECLARE_OP_USE_OUTPUT(ClipByValue)
DECLARE_OP_ADAPTER(AccumulateNV2)
DECLARE_OP_USE_DYN_INPUT(AccumulateNV2)
DECLARE_OP_USE_OUTPUT(AccumulateNV2)

13
mindspore/ccsrc/transform/graph_ir/op_declare/functional_ops_declare.cc
View File

@ -15,6 +15,7 @@
*/
#include "transform/graph_ir/op_declare/functional_ops_declare.h"
#include <string>
namespace mindspore::transform {
// Case
@ -31,4 +32,16 @@ ATTR_MAP(While) = {{"parallel_iterations", ATTR_DESC(parallel_iterations, AnyTra
DYN_OUTPUT_MAP(While) = {{0, DYN_OUTPUT_DESC(output)}};
SUBGRAPH_MAP(While) = {{0, SUBGRAPH_DESC(cond)}, {1, SUBGRAPH_DESC(body)}};
REG_ADPT_DESC(While, kNameWhile, ADPT_DESC(While));
// PartitionedCall
INPUT_MAP(PartitionedCall) = EMPTY_INPUT_MAP;
DYN_INPUT_MAP(PartitionedCall) = {{1, DYN_INPUT_DESC(args)}};
ATTR_MAP(PartitionedCall) = {
{"config", ATTR_DESC(config, AnyTraits<std::string>())},
{"config_proto", ATTR_DESC(config_proto, AnyTraits<std::string>())},
{"executor_type", ATTR_DESC(executor_type, AnyTraits<std::string>())},
};
DYN_OUTPUT_MAP(PartitionedCall) = {{0, DYN_OUTPUT_DESC(output)}};
SUBGRAPH_MAP(PartitionedCall) = {{0, SUBGRAPH_DESC(f)}};
REG_ADPT_DESC(PartitionedCall, kNamePartitionedCall, ADPT_DESC(PartitionedCall))
} // namespace mindspore::transform

5
mindspore/ccsrc/transform/graph_ir/op_declare/functional_ops_declare.h
View File

@ -32,5 +32,10 @@ DECLARE_OP_ATTR(While)
DECLARE_OP_USE_DYN_INPUT(While)
DECLARE_OP_USE_SUBGRAPH(While)
DECLARE_OP_USE_DYN_OUTPUT(While)
DECLARE_OP_ADAPTER(PartitionedCall)
DECLARE_OP_USE_DYN_INPUT(PartitionedCall)
DECLARE_OP_USE_SUBGRAPH(PartitionedCall)
DECLARE_OP_USE_DYN_OUTPUT(PartitionedCall)
} // namespace mindspore::transform
#endif // MINDSPORE_CCSRC_TRANSFORM_GRAPH_IR_OP_DECLARE_FUNCTIONAL_OPS_DECLARE_H_

8
mindspore/ccsrc/transform/graph_ir/op_declare/image_ops_declare.cc
View File

@ -60,4 +60,12 @@ ATTR_MAP(CropAndResize) = {{"extrapolation_value", ATTR_DESC(extrapolation_value
{"method", ATTR_DESC(method, AnyTraits<std::string>())}};
OUTPUT_MAP(CropAndResize) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(CropAndResize, kNameCropAndResize, ADPT_DESC(CropAndResize))
// DecodeImage
INPUT_MAP(DecodeImage) = {{1, INPUT_DESC(contents)}};
ATTR_MAP(DecodeImage) = {{"channels", ATTR_DESC(channels, AnyTraits<int64_t>())},
{"dtype", ATTR_DESC(dtype, AnyTraits<GEType>())},
{"expand_animations", ATTR_DESC(expand_animations, AnyTraits<bool>())}};
OUTPUT_MAP(DecodeImage) = {{0, OUTPUT_DESC(image)}};
REG_ADPT_DESC(DecodeImage, kNameDecodeImage, ADPT_DESC(DecodeImage))
} // namespace mindspore::transform

3
mindspore/ccsrc/transform/graph_ir/op_declare/image_ops_declare.h
View File

@ -39,5 +39,8 @@ DECLARE_OP_USE_OUTPUT(ResizeBilinearV2Grad)
DECLARE_OP_ADAPTER(CropAndResize)
DECLARE_OP_USE_OUTPUT(CropAndResize)
DECLARE_OP_ADAPTER(DecodeImage)
DECLARE_OP_USE_OUTPUT(DecodeImage)
} // namespace mindspore::transform
#endif // MINDSPORE_CCSRC_TRANSFORM_GRAPH_IR_OP_DECLARE_IMAGE_OPS_DECLARE_H_

4
mindspore/ccsrc/transform/graph_ir/op_declare/matrix_calculation_ops_declare.cc
View File

@ -96,6 +96,7 @@ ATTR_MAP(MatMulV2) = {{"transpose_a", ATTR_DESC(transpose_x1, AnyTraits<bool>())
{"transpose_b", ATTR_DESC(transpose_x2, AnyTraits<bool>())}};
OUTPUT_MAP(MatMulV2) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(MatMulV2, prim::kPrimMatMul->name(), ADPT_DESC(MatMulV2))
REG_ADPT_DESC(MatMulV2Duplicate, prim::kPrimMatMulV2->name(), ADPT_DESC(MatMulV2))
// MatrixDiag
INPUT_MAP(MatrixDiag) = {{1, INPUT_DESC(x)}};
@ -145,7 +146,8 @@ REG_ADPT_DESC(L2Loss, kNameL2Loss, ADPT_DESC(L2Loss))
INPUT_MAP(ScatterElements) = {{1, INPUT_DESC(data)}, {2, INPUT_DESC(indices)}, {3, INPUT_DESC(updates)}};
ATTR_MAP(ScatterElements) = {{"axis", ATTR_DESC(axis, AnyTraits<int64_t>())}};
OUTPUT_MAP(ScatterElements) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(ScatterElements, kNameTensorScatterElements, ADPT_DESC(ScatterElements))
REG_ADPT_DESC(TensorScatterElements, kNameTensorScatterElements, ADPT_DESC(ScatterElements))
REG_ADPT_DESC(ScatterElements, kNameScatterElements, ADPT_DESC(ScatterElements))
// FullyConnection
INPUT_MAP(FullyConnection) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(w)}, {3, INPUT_DESC(b)}, {4, INPUT_DESC(offset_w)}};

96
mindspore/ccsrc/transform/graph_ir/op_declare/ocr_ops_declare.cc Normal file
View File

@ -0,0 +1,96 @@
/**
* Copyright 2022 Huawei Technologies Co., Ltd
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "transform/graph_ir/op_declare/ocr_ops_declare.h"
#include <vector>
namespace mindspore::transform {
INPUT_MAP(OCRDetectionPreHandle) = {{1, INPUT_DESC(img)}};
ATTR_MAP(OCRDetectionPreHandle) = {{"format", ATTR_DESC(data_format, AnyTraits<std::string>())}};
OUTPUT_MAP(OCRDetectionPreHandle) = {
{0, OUTPUT_DESC(resized_img)}, {1, OUTPUT_DESC(h_scale)}, {2, OUTPUT_DESC(w_scale)}};
REG_ADPT_DESC(OCRDetectionPreHandle, kNameOCRDetectionPreHandle, ADPT_DESC(OCRDetectionPreHandle))
INPUT_MAP(OCRFindContours) = {{1, INPUT_DESC(img)}};
ATTR_MAP(OCRFindContours) = {{"value_mode", ATTR_DESC(value_mode, AnyTraits<int64_t>())}};
OUTPUT_MAP(OCRFindContours) = {
{0, OUTPUT_DESC(polys_data)}, {1, OUTPUT_DESC(polys_offset)}, {2, OUTPUT_DESC(polys_size)}};
REG_ADPT_DESC(OCRFindContours, kNameOCRFindContours, ADPT_DESC(OCRFindContours))
INPUT_MAP(BatchDilatePolys) = {{1, INPUT_DESC(polys_data)}, {2, INPUT_DESC(polys_offset)},
{3, INPUT_DESC(polys_size)}, {4, INPUT_DESC(score)},
{5, INPUT_DESC(min_border)}, {6, INPUT_DESC(min_area_thr)},
{7, INPUT_DESC(score_thr)}, {8, INPUT_DESC(expands_cale)}};
ATTR_MAP(BatchDilatePolys) = EMPTY_ATTR_MAP;
OUTPUT_MAP(BatchDilatePolys) = {
{0, OUTPUT_DESC(dilated_polys_data)}, {1, OUTPUT_DESC(dilated_polys_offset)}, {2, OUTPUT_DESC(dilated_polys_size)}};
REG_ADPT_DESC(BatchDilatePolys, kNameBatchDilatePolys, ADPT_DESC(BatchDilatePolys))
INPUT_MAP(ResizeAndClipPolys) = {
{1, INPUT_DESC(polys_data)}, {2, INPUT_DESC(polys_offset)}, {3, INPUT_DESC(polys_size)}, {4, INPUT_DESC(h_scale)},
{5, INPUT_DESC(w_scale)}, {6, INPUT_DESC(img_h)}, {7, INPUT_DESC(img_w)}};
ATTR_MAP(ResizeAndClipPolys) = EMPTY_ATTR_MAP;
OUTPUT_MAP(ResizeAndClipPolys) = {{0, OUTPUT_DESC(clipped_polys_data)},
{1, OUTPUT_DESC(clipped_polys_offset)},
{2, OUTPUT_DESC(clipped_polys_size)},
{3, OUTPUT_DESC(clipped_polys_num)}};
REG_ADPT_DESC(ResizeAndClipPolys, kNameResizeAndClipPolys, ADPT_DESC(ResizeAndClipPolys))
INPUT_MAP(OCRDetectionPostHandle) = {
{1, INPUT_DESC(img)}, {2, INPUT_DESC(polys_data)}, {3, INPUT_DESC(polys_offset)}, {4, INPUT_DESC(polys_size)}};
ATTR_MAP(OCRDetectionPostHandle) = {{"format", ATTR_DESC(data_format, AnyTraits<std::string>())}};
OUTPUT_MAP(OCRDetectionPostHandle) = {{0, OUTPUT_DESC(imgs_data)},
{1, OUTPUT_DESC(imgs_offset)},
{2, OUTPUT_DESC(imgs_size)},
{3, OUTPUT_DESC(rect_points)}};
REG_ADPT_DESC(OCRDetectionPostHandle, kNameOCRDetectionPostHandle, ADPT_DESC(OCRDetectionPostHandle))
INPUT_MAP(OCRIdentifyPreHandle) = {
{1, INPUT_DESC(imgs_data)}, {2, INPUT_DESC(imgs_offset)}, {3, INPUT_DESC(imgs_size)}};
ATTR_MAP(OCRIdentifyPreHandle) = {{"size", ATTR_DESC(size, AnyTraits<std::vector<int64_t>>())},
{"format", ATTR_DESC(data_format, AnyTraits<std::string>())}};
OUTPUT_MAP(OCRIdentifyPreHandle) = {{0, OUTPUT_DESC(resized_imgs)}};
REG_ADPT_DESC(OCRIdentifyPreHandle, kNameOCRIdentifyPreHandle, ADPT_DESC(OCRIdentifyPreHandle))
INPUT_MAP(BatchEnqueue) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(queue_id)}};
ATTR_MAP(BatchEnqueue) = {{"batch_size", ATTR_DESC(batch_size, AnyTraits<int64_t>())},
{"queue_name", ATTR_DESC(queue_name, AnyTraits<std::string>())},
{"queue_depth", ATTR_DESC(queue_depth, AnyTraits<int64_t>())},
{"pad_mode", ATTR_DESC(pad_mode, AnyTraits<std::string>())}};
OUTPUT_MAP(BatchEnqueue) = {{0, OUTPUT_DESC(enqueue_count)}};
REG_ADPT_DESC(BatchEnqueue, kNameBatchEnqueue, ADPT_DESC(BatchEnqueue))
INPUT_MAP(Dequeue) = {{1, INPUT_DESC(queue_id)}};
ATTR_MAP(Dequeue) = {{"output_type", ATTR_DESC(output_type, AnyTraits<GEType>())},
{"output_shape", ATTR_DESC(output_shape, AnyTraits<std::vector<int64_t>>())},
{"queue_name", ATTR_DESC(queue_name, AnyTraits<std::string>())}};
OUTPUT_MAP(Dequeue) = {{0, OUTPUT_DESC(data)}};
REG_ADPT_DESC(Dequeue, kNameDequeue, ADPT_DESC(Dequeue))
INPUT_MAP(OCRRecognitionPreHandle) = {{1, INPUT_DESC(imgs_data)},
{2, INPUT_DESC(imgs_offset)},
{3, INPUT_DESC(imgs_size)},
{4, INPUT_DESC(langs)},
{5, INPUT_DESC(langs_score)}};
ATTR_MAP(OCRRecognitionPreHandle) = {{"batch_size", ATTR_DESC(batch_size, AnyTraits<int64_t>())},
{"format", ATTR_DESC(data_format, AnyTraits<std::string>())},
{"pad_mode", ATTR_DESC(pad_mode, AnyTraits<std::string>())}};
OUTPUT_MAP(OCRRecognitionPreHandle) = {{0, OUTPUT_DESC(imgs)},
{1, OUTPUT_DESC(imgs_relation)},
{2, OUTPUT_DESC(imgs_lang)},
{3, OUTPUT_DESC(imgs_piece_fillers)}};
REG_ADPT_DESC(OCRRecognitionPreHandle, kNameOCRRecognitionPreHandle, ADPT_DESC(OCRRecognitionPreHandle))
} // namespace mindspore::transform

53
mindspore/ccsrc/transform/graph_ir/op_declare/ocr_ops_declare.h Normal file
View File

@ -0,0 +1,53 @@
/**
* Copyright 2022 Huawei Technologies Co., Ltd
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef MINDSPORE_CCSRC_TRANSFORM_GRAPH_IR_OP_DECLARE_OCR_OPS_DECLARE_H_
#define MINDSPORE_CCSRC_TRANSFORM_GRAPH_IR_OP_DECLARE_OCR_OPS_DECLARE_H_
#include <string>
#include <unordered_map>
#include "transform/graph_ir/op_declare/op_declare_macro.h"
#include "ops/ocr_ops.h"
namespace mindspore::transform {
DECLARE_OP_ADAPTER(OCRDetectionPreHandle)
DECLARE_OP_USE_OUTPUT(OCRDetectionPreHandle)
DECLARE_OP_ADAPTER(OCRFindContours)
DECLARE_OP_USE_OUTPUT(OCRFindContours)
DECLARE_OP_ADAPTER(BatchDilatePolys)
DECLARE_OP_USE_OUTPUT(BatchDilatePolys)
DECLARE_OP_ADAPTER(ResizeAndClipPolys)
DECLARE_OP_USE_OUTPUT(ResizeAndClipPolys)
DECLARE_OP_ADAPTER(OCRDetectionPostHandle)
DECLARE_OP_USE_OUTPUT(OCRDetectionPostHandle)
DECLARE_OP_ADAPTER(OCRIdentifyPreHandle)
DECLARE_OP_USE_OUTPUT(OCRIdentifyPreHandle)
DECLARE_OP_ADAPTER(BatchEnqueue)
DECLARE_OP_USE_OUTPUT(BatchEnqueue)
DECLARE_OP_ADAPTER(Dequeue)
DECLARE_OP_USE_OUTPUT(Dequeue)
DECLARE_OP_ADAPTER(OCRRecognitionPreHandle)
DECLARE_OP_USE_OUTPUT(OCRRecognitionPreHandle)
} // namespace mindspore::transform
#endif // MINDSPORE_CCSRC_TRANSFORM_GRAPH_IR_OP_DECLARE_OCR_OPS_DECLARE_H_

6
mindspore/ccsrc/transform/graph_ir/op_declare/reduce_ops_declare.cc
View File

@ -89,6 +89,7 @@ INPUT_MAP(ReduceMean) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(axes)}};
ATTR_MAP(ReduceMean) = {{"keep_dims", ATTR_DESC(keep_dims, AnyTraits<bool>())}};
OUTPUT_MAP(ReduceMean) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(ReduceMean, prim::kPrimReduceMean->name(), ADPT_DESC(ReduceMean))
REG_ADPT_DESC(ReduceMeanV1, "ReduceMeanV1", ADPT_DESC(ReduceMean))
// ReduceMinD
INPUT_MAP(ReduceMinD) = {{1, INPUT_DESC(x)}};
@ -105,4 +106,9 @@ INPUT_ATTR_MAP(ReduceMaxD) = {
ATTR_MAP(ReduceMaxD) = {{"keep_dims", ATTR_DESC(keep_dims, AnyTraits<bool>())}};
OUTPUT_MAP(ReduceMaxD) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(ReduceMaxD, prim::kPrimReduceMax->name(), ADPT_DESC(ReduceMaxD))
INPUT_MAP(ReduceMax) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(axes)}};
ATTR_MAP(ReduceMax) = {{"keep_dims", ATTR_DESC(keep_dims, AnyTraits<bool>())}};
OUTPUT_MAP(ReduceMax) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(ReduceMax, "ReduceMaxV1", ADPT_DESC(ReduceMax))
} // namespace mindspore::transform

3
mindspore/ccsrc/transform/graph_ir/op_declare/reduce_ops_declare.h
View File

@ -33,6 +33,9 @@ DECLARE_OP_ADAPTER(ReduceMaxD)
DECLARE_OP_USE_INPUT_ATTR(ReduceMaxD)
DECLARE_OP_USE_OUTPUT(ReduceMaxD)
DECLARE_OP_ADAPTER(ReduceMax)
DECLARE_OP_USE_OUTPUT(ReduceMax)
DECLARE_OP_ADAPTER(ReduceAllD)
DECLARE_OP_USE_INPUT_ATTR(ReduceAllD)
DECLARE_OP_USE_OUTPUT(ReduceAllD)

26
mindspore/ccsrc/transform/graph_ir/op_declare/rnn_declare.cc
View File

@ -16,6 +16,7 @@
#include "transform/graph_ir/op_declare/rnn_declare.h"
#include <string>
#include <vector>
namespace mindspore::transform {
// BasicLSTMCell
@ -139,4 +140,29 @@ ATTR_MAP(DynamicGRUV2Grad) = {{"direction", ATTR_DESC(direction, AnyTraits<std::
OUTPUT_MAP(DynamicGRUV2Grad) = {{0, OUTPUT_DESC(dw_input)}, {1, OUTPUT_DESC(dw_hidden)}, {2, OUTPUT_DESC(db_input)},
{3, OUTPUT_DESC(db_hidden)}, {4, OUTPUT_DESC(dx)}, {5, OUTPUT_DESC(dh_prev)}};
REG_ADPT_DESC(DynamicGRUV2Grad, kNameDynamicGRUV2Grad, ADPT_DESC(DynamicGRUV2Grad))
// CommonLSTM
INPUT_MAP(CommonLSTM) = {{1, INPUT_DESC(x)},
{2, INPUT_DESC(w)},
{3, INPUT_DESC(r)},
{4, INPUT_DESC(b)},
{5, INPUT_DESC(sequence_lens)},
{6, INPUT_DESC(initial_h)},
{7, INPUT_DESC(initial_c)},
{8, INPUT_DESC(p)}};
ATTR_MAP(CommonLSTM) = {
{"activation_alpha", ATTR_DESC(activation_alpha, AnyTraits<std::vector<float>>())},
{"activation_beta", ATTR_DESC(activation_beta, AnyTraits<std::vector<float>>())},
{"activations", ATTR_DESC(activations, AnyTraits<std::vector<std::string>>())},
{"clip", ATTR_DESC(clip, AnyTraits<float>())},
{"direction", ATTR_DESC(direction, AnyTraits<std::string>())},
{"hidden_size", ATTR_DESC(hidden_size, AnyTraits<int64_t>())},
{"input_forget", ATTR_DESC(input_forget, AnyTraits<int64_t>())},
};
OUTPUT_MAP(CommonLSTM) = {
{0, OUTPUT_DESC(y)},
{1, OUTPUT_DESC(y_h)},
{2, OUTPUT_DESC(y_c)},
};
REG_ADPT_DESC(CommonLSTM, "CommonLSTM", ADPT_DESC(CommonLSTM))
} // namespace mindspore::transform

3
mindspore/ccsrc/transform/graph_ir/op_declare/rnn_declare.h
View File

@ -48,5 +48,8 @@ DECLARE_OP_USE_OUTPUT(DynamicGRUV2)
DECLARE_OP_ADAPTER(DynamicGRUV2Grad)
DECLARE_OP_USE_OUTPUT(DynamicGRUV2Grad)
DECLARE_OP_ADAPTER(CommonLSTM)
DECLARE_OP_USE_OUTPUT(CommonLSTM)
} // namespace mindspore::transform
#endif // MINDSPORE_CCSRC_TRANSFORM_GRAPH_IR_OP_DECLARE_RNN_DECLARE_H_

12
mindspore/ccsrc/transform/graph_ir/op_declare/selection_ops_declare.cc
View File

@ -30,6 +30,7 @@ REG_ADPT_DESC(CumsumD, kNameCumSum, ADPT_DESC(CumsumD))
INPUT_MAP(GatherV2) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(indices)}, {3, INPUT_DESC(axis)}};
ATTR_MAP(GatherV2) = EMPTY_ATTR_MAP;
OUTPUT_MAP(GatherV2) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(GatherV2, prim::kPrimGatherV2->name(), ADPT_DESC(GatherV2))
// CumprodD
INPUT_MAP(CumprodD) = {{1, INPUT_DESC(x)}};
@ -39,6 +40,11 @@ ATTR_MAP(CumprodD) = {{"exclusive", ATTR_DESC(exclusive, AnyTraits<bool>())},
OUTPUT_MAP(CumprodD) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(CumprodD, kNameCumProd, ADPT_DESC(CumprodD))
INPUT_MAP(Tile) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(multiples)}};
ATTR_MAP(Tile) = EMPTY_ATTR_MAP;
OUTPUT_MAP(Tile) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(Tile, "TileV1", ADPT_DESC(Tile))
INPUT_MAP(Slice) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(offsets)}, {3, INPUT_DESC(size)}};
ATTR_MAP(Slice) = EMPTY_ATTR_MAP;
OUTPUT_MAP(Slice) = {{0, OUTPUT_DESC(y)}};
@ -109,6 +115,12 @@ ATTR_MAP(RangeD) = {{"start", ATTR_DESC(start, AnyTraits<float>())},
OUTPUT_MAP(RangeD) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(RangeD, kNameRange, ADPT_DESC(RangeD))
// RangeV2
INPUT_MAP(Range) = {{1, INPUT_DESC(start)}, {2, INPUT_DESC(limit)}, {3, INPUT_DESC(delta)}};
ATTR_MAP(Range) = EMPTY_ATTR_MAP;
OUTPUT_MAP(Range) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(RangeV2, kNameRangeV2, ADPT_DESC(Range))
// InplaceAddD
INPUT_MAP(InplaceAddD) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(v)}};
ATTR_MAP(InplaceAddD) = {{"indices", ATTR_DESC(indices, AnyTraits<std::vector<int64_t>>())}};

6
mindspore/ccsrc/transform/graph_ir/op_declare/selection_ops_declare.h
View File

@ -73,6 +73,9 @@ DECLARE_OP_ADAPTER(CumprodD)
DECLARE_OP_USE_INPUT_ATTR(CumprodD)
DECLARE_OP_USE_OUTPUT(CumprodD)
DECLARE_OP_ADAPTER(Tile)
DECLARE_OP_USE_OUTPUT(Tile)
DECLARE_OP_ADAPTER(TileD)
DECLARE_OP_USE_INPUT_ATTR(TileD)
DECLARE_OP_USE_OUTPUT(TileD)
@ -87,6 +90,9 @@ DECLARE_OP_USE_OUTPUT(GatherV2D)
DECLARE_OP_ADAPTER(RangeD)
DECLARE_OP_USE_OUTPUT(RangeD)
DECLARE_OP_ADAPTER(Range)
DECLARE_OP_USE_OUTPUT(Range)
DECLARE_OP_ADAPTER(InplaceAddD)
DECLARE_OP_USE_OUTPUT(InplaceAddD)

11
mindspore/ccsrc/transform/graph_ir/op_declare/string_ops_declare.cc
View File

@ -16,11 +16,20 @@
#include "transform/graph_ir/op_declare/string_ops_declare.h"
#include <vector>
#include <string>
namespace mindspore::transform {
INPUT_MAP(StringUpper) = {{1, INPUT_DESC(input)}};
ATTR_MAP(StringUpper) = {{"encoding", ATTR_DESC(encoding, AnyTraits<std::string>())}};
OUTPUT_MAP(StringUpper) = {{0, OUTPUT_DESC(output)}};
REG_ADPT_DESC(StringUpper, kNameStringUpper, ADPT_DESC(StringUpper))
INPUT_MAP(StringLength) = {{1, INPUT_DESC(x)}};
ATTR_MAP(StringLength) = {{"unit", ATTR_DESC(unit, AnyTraits<std::string>())}};
OUTPUT_MAP(StringLength) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(StringLength, kNameStringLength, ADPT_DESC(StringLength))
INPUT_MAP(DecodeBase64) = {{1, INPUT_DESC(x)}};
ATTR_MAP(DecodeBase64) = EMPTY_ATTR_MAP;
OUTPUT_MAP(DecodeBase64) = {{0, OUTPUT_DESC(y)}};
REG_ADPT_DESC(DecodeBase64, kNameDecodeBase64, ADPT_DESC(DecodeBase64))
} // namespace mindspore::transform

8
mindspore/ccsrc/transform/graph_ir/op_declare/string_ops_declare.h
View File

@ -17,11 +17,19 @@
#ifndef MINDSPORE_CCSRC_TRANSFORM_GRAPH_IR_OP_DECLARE_STRING_OPS_DECLARE_H_
#define MINDSPORE_CCSRC_TRANSFORM_GRAPH_IR_OP_DECLARE_STRING_OPS_DECLARE_H_
#include <string>
#include <unordered_map>
#include "transform/graph_ir/op_declare/op_declare_macro.h"
#include "ops/string_ops.h"
namespace mindspore::transform {
DECLARE_OP_ADAPTER(StringUpper)
DECLARE_OP_USE_OUTPUT(StringUpper)
DECLARE_OP_ADAPTER(StringLength)
DECLARE_OP_USE_OUTPUT(StringLength)
DECLARE_OP_ADAPTER(DecodeBase64)
DECLARE_OP_USE_OUTPUT(DecodeBase64)
} // namespace mindspore::transform
#endif // MINDSPORE_CCSRC_TRANSFORM_GRAPH_IR_OP_DECLARE_STRING_OPS_DECLARE_H_

6
mindspore/ccsrc/transform/graph_ir/op_declare/transformation_ops_declare.cc
View File

@ -48,6 +48,12 @@ ATTR_MAP(TransposeD) = EMPTY_ATTR_MAP;
// Do not set Transpose operator output descriptor
REG_ADPT_DESC(TransposeD, prim::kPrimTranspose->name(), ADPT_DESC(TransposeD))
// Transpose
INPUT_MAP(Transpose) = {{1, INPUT_DESC(x)}, {2, INPUT_DESC(perm)}};
ATTR_MAP(Transpose) = EMPTY_ATTR_MAP;
// Do not set Transpose operator output descriptor
REG_ADPT_DESC(Transpose, "TransposeV1", ADPT_DESC(Transpose))
// SpaceToDepth
INPUT_MAP(SpaceToDepth) = {{1, INPUT_DESC(x)}};
ATTR_MAP(SpaceToDepth) = {{"block_size", ATTR_DESC(block_size, AnyTraits<int64_t>())}};

2
mindspore/ccsrc/transform/graph_ir/op_declare/transformation_ops_declare.h
View File

@ -31,6 +31,8 @@ DECLARE_OP_USE_DYN_OUTPUT(Unpack)
DECLARE_OP_ADAPTER(TransposeD)
DECLARE_OP_USE_INPUT_ATTR(TransposeD)
DECLARE_OP_ADAPTER(Transpose)
DECLARE_OP_ADAPTER(Flatten)
DECLARE_OP_USE_OUTPUT(Flatten)

5
mindspore/core/ops/core_ops.h
View File

@ -568,6 +568,7 @@ GVAR_DEF(PrimitivePtr, kPrimScatterNdMin, std::make_shared<Primitive>("ScatterNd
GVAR_DEF(PrimitivePtr, kPrimScatterNdMul, std::make_shared<Primitive>("ScatterNdMul"));
GVAR_DEF(PrimitivePtr, kPrimScatterNdDiv, std::make_shared<Primitive>("ScatterNdDiv"));
GVAR_DEF(PrimitivePtr, kPrimScatterUpdate, std::make_shared<Primitive>("ScatterUpdate"));
GVAR_DEF(PrimitivePtr, kPrimScatterElements, std::make_shared<Primitive>("ScatterElements"));
GVAR_DEF(PrimitivePtr, kPrimScatterAddWithAxis, std::make_shared<Primitive>(kScatterAddWithAxis));
GVAR_DEF(PrimitivePtr, kPrimTensorScatterElements, std::make_shared<Primitive>("TensorScatterElements"));
GVAR_DEF(PrimitivePtr, kPrimTensorScatterUpdate, std::make_shared<Primitive>("TensorScatterUpdate"));
@ -1443,11 +1444,15 @@ GVAR_DEF(PrimitivePtr, kPrimTensorArrayStack, std::make_shared<Primitive>("Tenso
GVAR_DEF(PrimitivePtr, kPrimTensorArray, std::make_shared<Primitive>("TensorArray"));
GVAR_DEF(PrimitivePtr, kPrimTensorArrayWrite, std::make_shared<Primitive>("TensorArrayWrite"));
GVAR_DEF(PrimitivePtr, kPrimTensorArrayGather, std::make_shared<Primitive>("TensorArrayGather"));
GVAR_DEF(PrimitivePtr, kPrimPartitionedCall, std::make_shared<Primitive>("PartitionedCall"));
GVAR_DEF(PrimitivePtr, kPrimDecodeImage, std::make_shared<Primitive>("DecodeImage"));
GVAR_DEF(PrimitivePtr, kPrimStridedSliceV2, std::make_shared<Primitive>("StridedSliceV2"));
GVAR_DEF(PrimitivePtr, kPrimKMeansCentroids, std::make_shared<Primitive>("KMeansCentroids"));
GVAR_DEF(PrimitivePtr, kPrimReservoirReplayBufferCreate, std::make_shared<Primitive>("ReservoirReplayBufferCreate"));
GVAR_DEF(PrimitivePtr, kPrimReservoirReplayBufferPush, std::make_shared<Primitive>("ReservoirReplayBufferPush"));
GVAR_DEF(PrimitivePtr, kPrimReservoirReplayBufferSample, std::make_shared<Primitive>("ReservoirReplayBufferSample"));
GVAR_DEF(PrimitivePtr, kPrimReservoirReplayBufferDestroy, std::make_shared<Primitive>("ReservoirReplayBufferDestroy"));
GVAR_DEF(PrimitivePtr, kPrimOCRDetectionPreHandle, std::make_shared<Primitive>("OCRDetectionPreHandle"));
// AdamApplyOne
GVAR_DEF(PrimitivePtr, kPrimAdamApplyOne, std::make_shared<Primitive>("AdamApplyOne"));

53
mindspore/core/ops/scatter_elements.cc Normal file
View File

@ -0,0 +1,53 @@
/**
* Copyright 2022 Huawei Technologies Co., Ltd
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "ops/scatter_elements.h"
#include <set>
#include <string>
#include "ops/op_utils.h"
#include "utils/check_convert_utils.h"
#include "abstract/ops/primitive_infer_map.h"
#include "mindapi/src/helper.h"
namespace {
constexpr size_t kScatterElementsArgSize = 3;
} // namespace
namespace mindspore {
namespace ops {
int64_t ScatterElements::get_axis() const { return axis_; }
void ScatterElements::set_axis(const int64_t axis) { axis_ = axis; }
AbstractBasePtr ScatterElementsInfer(const abstract::AnalysisEnginePtr &, const PrimitivePtr &primitive,
const abstract::AbstractBasePtrList &args_spec_list) {
const std::string op_name = primitive->name();
CheckRequiredArgsSize(op_name, args_spec_list, kScatterElementsArgSize);
auto x = abstract::CheckArg<abstract::AbstractTensor>(op_name, args_spec_list, 0);
MS_EXCEPTION_IF_NULL(x);
MS_EXCEPTION_IF_NULL(x->shape());
ShapeVector shape = x->shape()->shape();
ShapeVector min_shape = x->shape()->min_shape();
ShapeVector max_shape = x->shape()->max_shape();
abstract::CheckMinMaxShape(shape, &min_shape, &max_shape);
return std::make_shared<abstract::AbstractTensor>(x->element(),
std::make_shared<abstract::Shape>(shape, min_shape, max_shape));
}
MIND_API_OPERATOR_IMPL(ScatterElements, BaseOperator);
REGISTER_PRIMITIVE_EVAL_IMPL(ScatterElements, prim::kPrimScatterElements, ScatterElementsInfer, nullptr, true);
} // namespace ops
} // namespace mindspore

48
mindspore/core/ops/scatter_elements.h Normal file
View File

@ -0,0 +1,48 @@
/**
* Copyright 2022 Huawei Technologies Co., Ltd
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef MINDSPORE_CORE_OPS_SCATTER_ELEMENTS_H_
#define MINDSPORE_CORE_OPS_SCATTER_ELEMENTS_H_
#include <vector>
#include <memory>
#include "ops/base_operator.h"
#include "mindapi/base/types.h"
namespace mindspore {
namespace ops {
constexpr auto kNameScatterElements = "ScatterElements";
class MIND_API ScatterElements : public BaseOperator {
public:
MIND_API_BASE_MEMBER(ScatterElements);
/// \brief Constructor.
ScatterElements() : BaseOperator(kNameScatterElements) { InitIOName({"indices", "update", "shape"}, {"output"}); }
/// \brief Init.
void Init() const {}
void set_axis(const int64_t axis);
int64_t get_axis() const;
private:
int64_t axis_ = 0;
};
abstract::AbstractBasePtr ScatterElementsInfer(const abstract::AnalysisEnginePtr &, const PrimitivePtr &primitive,
const std::vector<abstract::AbstractBasePtr> &input_args);
} // namespace ops
} // namespace mindspore
#endif // MINDSPORE_CORE_OPS_SCATTER_ELEMENTS_H_

4
mindspore/core/ops/tensor_scatter_elements.h
View File

@ -25,14 +25,14 @@
namespace mindspore {
namespace ops {
constexpr auto kNameScatterElements = "TensorScatterElements";
constexpr auto kNameTensorScatterElements = "TensorScatterElements";
/// \brief Updates tensor values by using input indices and value.
/// Refer to Python API @ref mindspore.ops.TensorScatterElements for more details.
class MIND_API TensorScatterElements : public BaseOperator {
public:
MIND_API_BASE_MEMBER(TensorScatterElements);
/// \brief Constructor.
TensorScatterElements() : BaseOperator(kNameScatterElements) {
TensorScatterElements() : BaseOperator(kNameTensorScatterElements) {
InitIOName({"input_x", "indices", "update"}, {"output"});
}
/// \brief Init. Refer to the parameters of Python API @ref mindspore.ops.TensorScatterElements for the inputs.

55
mindspore/python/mindspore/ops/operations/_inner_ops.py
View File

@ -1521,6 +1521,34 @@ class NonZeroWithValueShape(Primitive):
self.init_prim_io_names(inputs=['value', 'index', 'count'], outputs=['out_value', 'out_index'])
class DecodeImage(PrimitiveWithInfer):
"""
Returns image data that parse from string Tensor.
Inputs:
- **x** (Tensor), a Tensor of type string. 0-D. The jPEG, GIF, PNG, BMP-encoded image.
Outputs:
A Tensor of type uint8, uint16, float.
Supported Platforms:
``Ascend``
Examples:
"""
@prim_attr_register
def __init__(self, channels=0, dtype=mstype.uint8, expand_animations=False, _op_max_shape="8192,8192,3",
_op_max_size=[8000000]):
self.init_prim_io_names(inputs=["contents"], outputs=["image"])
self.res_type = dtype
def infer_shape(self, x):
return (-1, -1, 3)
def infer_dtype(self, x):
return self.res_type
class SliceGetItem(Primitive):
"""
using SliceGetItem to get slice's attribute of 'start' 'stop' 'step'
@ -1727,6 +1755,33 @@ class ParallelResizeBilinear(PrimitiveWithInfer):
'value': None}
class PartitionedCall(PrimitiveWithInfer):
"""
Pass the input tensors to the subgraph and return the output tensors.
Inputs:
- **inputs** (Tuple), the input tensors, which will be passed to subgraph.
Outputs:
- outputs(Tuple), the output tensor returned by subgraph.
Supported Platforms:
``Ascend``
Examples:
"""
@prim_attr_register
def __init__(self, graph, executor_type=""):
self.add_prim_attr("executor_type", executor_type)
self.graph = graph
def infer_shape(self, *inputs):
return NotImplementedError
def infer_dtype(self, *inputs):
return NotImplementedError
class CellBackwardHook(PrimitiveWithInfer):
r"""
This operator is used to hook input gradient and output gradient of Cell object.

2
mindspore/python/mindspore/ops/operations/_ocr_ops.py
View File

@ -35,7 +35,7 @@ class GetShape(PrimitiveWithInfer):
self.init_prim_io_names(inputs=["x"], outputs=["y"])
def infer_shape(self, x):
return (x[0],)
return (len(x[0]),)
def infer_dtype(self, x):
return mstype.int32

9
mindspore/python/mindspore/ops/operations/array_ops.py
View File

@ -900,6 +900,13 @@ class TensorShape(Primitive):
self.init_prim_io_names(inputs=['input_x'], outputs=['output'])
class Unsqueeze(PrimitiveWithCheck):
@prim_attr_register
def __init__(self, axis):
self.init_prim_io_names(inputs=['x'], outputs=['y'])
self.axis = axis
class DynamicShape(Primitive):
"""
Same as operator TensorShape. DynamicShape will be deprecated in the future.
@ -911,7 +918,7 @@ class DynamicShape(Primitive):
@deprecated("1.7", "TensorShape", True)
@prim_attr_register
def __init__(self):
def __init__(self, dtype=9):
"""init Shape"""
self.init_prim_io_names(inputs=['tensor'], outputs=['output'])
self.add_prim_attr('is_dynamic_shape', True)