cavis/libnd4j/blas/cpu/NDArrayFactory.cpp
Alex Black 1170827c18 Merge master to upstream (#7945)
* Shugeo strided slice zeros (#14)

* Modified strided_slice op to properly work with empty-like shapes.

* Fixed test for reduce_mean with empty-like input.

* [WIP] Last merge (#15)

* correct logsoftmax looss (#2)

* Small SameDiff listener fix (#4)

* Various fixes (#6)

* #7839 Fix for asXMatrix and tests

* #7866 EmbeddingSequenceLayer dtype fix + test

* #7856 SameDiff save/load stream methods

* #7859 RegressionEvaluation rank 4 fix + tests + axis configuration

* EvaluationBinary 3d/4d

* More evaluation 3d/4d tests

* #7847 Evaluation empty checks

* Small test ifx

* #7848 Fix median edge case

* Improve DL4J samediff layer tests

* [WIP] FastText wrapper implemented (#8)

* FastText implemented

* Some fixes

* Fix shapes for wordsNearest

* Validation of input vectors

* Fixes

* Fixed test

* Thread tagged

* Some tweaks

* setContextClassLoader for DeallocatorServiceThread

* Numpy format tests (#1)

* Various fixes (#11)

* #7852 SameDiff gather fix

* #7892 SameDiff placeholder to constant conversion

* #7890 validate input rank for MLN/CG init methods

* Fix broken permute shape calculation

* Permute and gather fixes

* Tests

* #7850 LogSumExp fix + test

* Handful of test fixes

* Empty arrays with non-scalar shapes (#10)

* minor rearrangements for lambdas

* empty tensors with non-scalar shapes

* numpy empty tensors with non-scalar shapes

* few more empty tweaks

* Small fixes

* conv3d signature update

* micro fix in batchnorm mkldnn

* Import fixes

* Fix

* MKL-DNN update

* Small fill fix

* fill with empty input + test

* Fixes

* Small error improvement

* Fix

* one special test

* couple of fixes for lstm

* Rewrite TFGraphMapper.getNDArrayFromTensor to be maintainable and less error prone

* Fixes

* FP16

* Unsigned

* BFloat16

* Fill op - empty tweaks

* - couple of fixes for empty arrays construction
- stack updated

* strided slice fix

* one transform test

* provide method for reducing shapeInfo in case of input array is empty

* Fixed reduceAlongDimensions to use empty input properly.

* couple of broadcast tests

* couple of tests broadcast tests + tweak to make them pass

* add check of non-empty to methods producing sub-arrays

* Fixed reshapeC with zeros in shape.

* complete empty check in reduce_... legacy ops

* Concat and cumsum/prod

* Tweak to empty shape inference on import

* add empty check to the rest of reduce legacy ops

* one more test

* correct typo in evalReduceShapeInfoEmpty

* Added tests for reduce_* ops to tests with zero shapes.

* few more tests for empty reductions

* Fixed strided_slice op with empty case and tests.

* one more empty reduction test

* Fixed strided_slice test.

* add empty check to NDArray::reshapei

* infOrMax

* empty min/max with infinity tests

* made unstack working correctly with empty arrays

* few IndexReduce tests + tweaks for empty shapes

* add test for empty concat

* few tests fixed

* Validation fix for reductions on empty shapes

* Reverse fix

* Reduction shape calc fixes

* SameDiff.generateOutputVariable: don't use shape function to determine number of outputs

* Range fix

* - NDArray constructor updated for scalars/empty arrays
- few tests fixed

* More fixes

* Empty creator fixes

* concat fix

* concat fix

* TF import tests: allow 'both all NaN' and 'both all inf' to pass

* Slice, zero fraction, and reshape fixes

* transpose, gather

* Zero fraction

* scalar cast fix

* Empty reduction axis support

* few more tests fixed

* Fixed input checks conforming with TF for concat op and tests.

* few tests fixed

* matmul scalar shape fix

* Fixed checkout for data type and scalarity with concat to allow non-empty scalars with vector concats.

* broadcast bool fix

* few more tests

* few more tests

* correct evalReduceShapeInfoEmpty

* argmax/argmin + tests

* one more empty edge case + one more test

* argmax/argmin/realdiv_bp tweaks

* empty reshape test + fix

* Helper fixes

* Small fixes

* Gather test fix

* Gather test fix

* Small fixes

* reduce scalar zero values

* scalar mean workaround

* Remove debug code

* along dim mean workaround

* one more test

* - equalsTo() tweak for empty arrays
- one more test

* broadcast tweaks

* [WIP] Fixing outstanding issues for NLP (#9)

* Avoid using not-inited objects

* Test fixed.

* Redundant method avoided for models like FastText

* KMeans++ implementation

* KMeans++ implementation

* Disable parallel execution

* KMeans++

* Tests

* Dev branch merge (#16)

* SameDiff: convertDataType and gradient check util improvements (#12)

* GradCheck util improvements

* StopGradient constructor + test

* SameDiff: Add datatype conversion

* Javadoc and add DataType.isNumerical()

* Small fix

* Fix SameDiff TF import test cases intermediate naming (workaround for bad default)

* TFGraphTestAllHelper: check intermediates in execution order

* Add missing debug listener

* [WIP] lstmBlock fix + other changes (#13)

- fixes lstmBlock issue
- changes NDArray method reshape(), permute(), transpose() by making them return instance instead of pointer
- CheckNumerics op
- fixes for ReduceBool IsInfOrNan & IsFinite

* Small test fix

* CheckNumerics op wrapper

* Fix some issues on master (#17)

* Fix DataVec test issue

* Fix issue with dl4j SameDiff output layer

* Dtype fix for lambda layers

* #7912 BertIterator dtype fix (use float32 not global default)

* [WIP] Next set of CUDA stuff (#7)

New CUDA implementations and improvements

* bad file

* Dev branch master merge (#23)

* SameDiff: convertDataType and gradient check util improvements (#12)

* GradCheck util improvements

* StopGradient constructor + test

* SameDiff: Add datatype conversion

* Javadoc and add DataType.isNumerical()

* Small fix

* Fix SameDiff TF import test cases intermediate naming (workaround for bad default)

* TFGraphTestAllHelper: check intermediates in execution order

* Add missing debug listener

* [WIP] lstmBlock fix + other changes (#13)

- fixes lstmBlock issue
- changes NDArray method reshape(), permute(), transpose() by making them return instance instead of pointer
- CheckNumerics op
- fixes for ReduceBool IsInfOrNan & IsFinite

* Small test fix

* CheckNumerics op wrapper

* Compatibility of deserialization (#18)

Signed-off-by: Alexander Stoyakin <alexander.stoyakin@gmail.com>

* SameDiff: add activation gradient checking support for debugging (#19)

* SameDiff gradient checker: first pass on activation gradient checks

* Fixes + tests for activation gradient checking

* Javadoc

* [WIP] Some nd4j data type corrections (#20)

* Adjust data type

* Set correct Data type.

* Size of proper data type.

* fix averaged cpu load (#22)

* SameDiff ops, TF import and fixes (#24)

* CheckNumerics tests + fixes + misc fixes

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Fake quant

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Fixes

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* FakeQuantWithMinMaxArgs

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* CheckNumerics fix

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Fix libnd4j ALL_INTS and ALL_FLOATS declaration (uint and bfloat types)

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Small fix

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Javadoc

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Exception tweak

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* fix

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Fix for out of scope stack allocated var use

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Ignores

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Ignore for known failing test (already logged issue)

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Merge upstream to fork (#25)

* Add thousand-separator commas to TotalParams (#7915)

* Add thousand-separator commas to TotalParams

The number of parameters can be quite large, and it would help the reading of the summary printout to have the TotalParams column & values at the bottom have thousand-separator-commas in them.

* Add thousand-separator commas to MultiLayerNetwork

Corresponding change to MultiLayerNetwork

Signed-off-by: Jxtps Jxtps <jxtps435@gmail.com>

* Update contributing and issue/PR templates (#7934)

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Fix link to AdaDelta paper (#7942)

Fix link to AdaDelta paper hosted on matthewzeiler.com

Signed-off-by: Jxtps

* Fixes, and ignores for known/logged failing issues (#7943)

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* SameDiff + DL4J/SameDiff: Multiple fixes (#28)

* #7919 HDF5 attribute buffer length fix

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* #7909 Arbiter constructor exception ux improvements

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* #7925 RNN output layer length checks

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* #7939 Add listener for validating inputs are not incorrectly modified

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* #7939 Integrate NonInplaceValidationListener into tests

* #7844 DL4J SameDiff fixes for variable minibatch size

* DL4J SameDiff fixes - ensure gradient for input placeholder is available

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Tweaks to ExternalErrorsFunction - use placeholders, make more robust

* Another fix

* More fixes

* More SameDiff/DL4J fixes

* Scope out scalar array creation in BaseScalarOp

* Remove debug code

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* [WIP] Final dev branch merge (#29)

* SameDiff: convertDataType and gradient check util improvements (#12)

* GradCheck util improvements

* StopGradient constructor + test

* SameDiff: Add datatype conversion

* Javadoc and add DataType.isNumerical()

* Small fix

* Fix SameDiff TF import test cases intermediate naming (workaround for bad default)

* TFGraphTestAllHelper: check intermediates in execution order

* Add missing debug listener

* [WIP] lstmBlock fix + other changes (#13)

- fixes lstmBlock issue
- changes NDArray method reshape(), permute(), transpose() by making them return instance instead of pointer
- CheckNumerics op
- fixes for ReduceBool IsInfOrNan & IsFinite

* Small test fix

* CheckNumerics op wrapper

* Compatibility of deserialization (#18)

Signed-off-by: Alexander Stoyakin <alexander.stoyakin@gmail.com>

* SameDiff: add activation gradient checking support for debugging (#19)

* SameDiff gradient checker: first pass on activation gradient checks

* Fixes + tests for activation gradient checking

* Javadoc

* [WIP] Some nd4j data type corrections (#20)

* Adjust data type

* Set correct Data type.

* Size of proper data type.

* fix averaged cpu load (#22)

* [WIP] Multiple dataset iterators (#27)

* Splitting dataset into arbitrary number

* Fixes

* Multiple split of iterator

* Test

* Test

* Some fixes

* signature change

* one more tweak

Signed-off-by: raver119 <raver119@gmail.com>

* one more test for sequential use of DataSetIteratorSplitter

Signed-off-by: raver119 <raver119@gmail.com>

* Fixes

* Fixes

* one more test for Alexander

Signed-off-by: raver119 <raver119@gmail.com>

* Some fixes

* Some fixes

* one more test for Alexander

Signed-off-by: raver119 <raver119@gmail.com>

* minor test fix

Signed-off-by: raver119 <raver119@gmail.com>

* Some fixes

* Some fixes

* couple of assertions tweaked

Signed-off-by: raver119 <raver119@gmail.com>

* MDS splitter test :/

Signed-off-by: raver119 <raver119@gmail.com>

* Minor refactoring

* Multi dataset

* Some fixes

* More tests

* Small number of test fixes/improvements (failures on CI) (#31)

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* [WIP] More CUDA stuff (#26)

* initial commit

Signed-off-by: raver119 <raver119@gmail.com>

* LRN BP CUDA

Signed-off-by: raver119 <raver119@gmail.com>

* less memory

Signed-off-by: raver119 <raver119@gmail.com>

* Fixed bug with crop_and_resize op helper.

* get rid of unnecessary index-calculation dunction

Signed-off-by: Yurii <yurii@skymind.io>

* Fixed sort with nth_element cuda-based helper.

* Refactored nth_element.

* Refactored nth_element op and tests.

* Modified usage of dim array with sortTad routine.

* Refactored main routine of helper for non_max_image_suppression op.

* non_max_image_suppression op helper with cuda kernel implementation. Initial revision.

* fix vol2col cuda kernel

* meh

Signed-off-by: raver119 <raver119@gmail.com>

* topK concept

Signed-off-by: raver119 <raver119@gmail.com>

* unsorted topK with scanWitdh of 1

Signed-off-by: raver119 <raver119@gmail.com>

* correct vol2col tests

* sorted/unsorted topK

Signed-off-by: raver119 <raver119@gmail.com>

* implementation and fixing col2im/col2vol

* Corrected usage flags with input/output with reverse op.

* dup is const now

Signed-off-by: raver119 <raver119@gmail.com>

* percentile op

Signed-off-by: raver119 <raver119@gmail.com>

* group tests for mapool2d

Signed-off-by: Yurii <yurii@skymind.io>

* special test for george

Signed-off-by: raver119 <raver119@gmail.com>

* less threads for sortTad

Signed-off-by: raver119 <raver119@gmail.com>

* provide conv2d for cuda

Signed-off-by: Yurii <yurii@skymind.io>

* remove auther in sort tad kernel code

Signed-off-by: Yurii <yurii@skymind.io>

* provide depthwise_conv2d for cuda

Signed-off-by: Yurii <yurii@skymind.io>

* - max_pooling_with_argmax
- null check for special use

Signed-off-by: raver119 <raver119@gmail.com>

* dts cuda

Signed-off-by: raver119 <raver119@gmail.com>

* provide sconv2d for cuda

Signed-off-by: Yurii <yurii@skymind.io>

* std cuda

Signed-off-by: raver119 <raver119@gmail.com>

* Refactored non_max_suppression op to conform TF implementation.

* Improved suppression helper.

* provide pooling3d for cuda

Signed-off-by: Yurii <yurii@skymind.io>

* minor lstm rearrangements

Signed-off-by: raver119 <raver119@gmail.com>

* more of minor lstm rearrangements

Signed-off-by: raver119 <raver119@gmail.com>

* (bi)dynamic_rnn

Signed-off-by: raver119 <raver119@gmail.com>

* templates init order

Signed-off-by: raver119 <raver119@gmail.com>

* Refactored non_max_suppression op.

* Added cuda kernel for non_max_suppression.

* CPU sort by key/value

Signed-off-by: raver119 <raver119@gmail.com>

* CPU sort TAD by key/value

Signed-off-by: raver119 <raver119@gmail.com>

* CPU sort TAD by key/value tests

Signed-off-by: raver119 <raver119@gmail.com>

* Eliminate compiler error with cuda implementation.

* - repaired gradCheck in cuda
- provide conv2d_bp for cuda

Signed-off-by: Yurii <yurii@skymind.io>

* missed signature

Signed-off-by: raver119 <raver119@gmail.com>

* provide depthwise_conv2d_bp for cuda

Signed-off-by: Yurii <yurii@skymind.io>

* Implementation of lup helper with cuda kernel. Initial commit.

* further work on backprops for convolutions

Signed-off-by: Yurii <yurii@skymind.io>

* CUDA linear sort by key/val

Signed-off-by: raver119 <raver119@gmail.com>

* CUDA tad sort by key/val

Signed-off-by: raver119 <raver119@gmail.com>

* start providing of backprop for pooling2d/3d

Signed-off-by: Yurii <yurii@skymind.io>

* Added atomicAdd for bool datatype.

* dynamic partition concept

Signed-off-by: raver119 <raver119@gmail.com>

* dynamic partition concept

Signed-off-by: raver119 <raver119@gmail.com>

* dynamic partition scalar CUDA

Signed-off-by: raver119 <raver119@gmail.com>

* important comment

Signed-off-by: raver119 <raver119@gmail.com>

* fix pooling2d/3d backprop helpers

Signed-off-by: Yurii <yurii@skymind.io>

* Added non-linear test with dynamic_partition.

* Improved test for dynamic_partition.

* dynamic_partition TAD concept

Signed-off-by: raver119 <raver119@gmail.com>

* - dynamic_partition TAD CUDA impl
- dynamic_partition TAD CPU fix

Signed-off-by: raver119 <raver119@gmail.com>

* - rewrite cpu code for usampling2d/3d
- write cuda code for usampling2d/3d

Signed-off-by: Yurii <yurii@skymind.io>

* dynamic_stitch CUDA vector case

Signed-off-by: raver119 <raver119@gmail.com>

* dynamic_stitch CUDA TAD case concept

Signed-off-by: raver119 <raver119@gmail.com>

* dynamic_stitch CUDA TAD case impl

Signed-off-by: raver119 <raver119@gmail.com>

* Added tests for dynamic_stitch 3D-4D cases.

* minor tests tweaks

Signed-off-by: raver119 <raver119@gmail.com>

* Fixed type check for dynamic stitch.

* min/max bp

Signed-off-by: raver119 <raver119@gmail.com>

* rewrite code for upsampling2d/3d cpu

Signed-off-by: Yurii <yurii@skymind.io>

* reduce min/max/norm_max bp

Signed-off-by: raver119 <raver119@gmail.com>

* lup implementation. Additional enhancements.

* provide code for upsamling2d/3d backprop

Signed-off-by: Yurii <yurii@skymind.io>

* weightedCrossEntropyWithLogits

Signed-off-by: raver119 <raver119@gmail.com>

* Fixed template math atomicMul for 64bit ints.

* Refactored dynamic_partition_bp op.

* inverseBroadcast fix

Signed-off-by: raver119 <raver119@gmail.com>

* DynamicPartitionBP test datatype fixed.

* - nd4j_atomicMul Windows fix
- cpu/NDArrayLambda.hpp excluded from CUDA

Signed-off-by: raver119 <raver119@gmail.com>
2019-06-27 18:37:04 +03:00

621 lines
36 KiB
C++

/*******************************************************************************
* Copyright (c) 2015-2018 Skymind, Inc.
*
* This program and the accompanying materials are made available under the
* terms of the Apache License, Version 2.0 which is available at
* https://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.
*
* SPDX-License-Identifier: Apache-2.0
******************************************************************************/
//
// Created by GS <sgazeos@gmail.com> on 2018-12-20.
//
#include <NDArrayFactory.h>
#include <exceptions/cuda_exception.h>
#include <ConstantHelper.h>
#include <ConstantShapeHelper.h>
#include <ShapeUtils.h>
#include <type_traits>
namespace nd4j {
////////////////////////////////////////////////////////////////////////
template <>
NDArray NDArrayFactory::create<bool>(const char order, const std::vector<Nd4jLong> &shape, const std::vector<bool> &data, nd4j::LaunchContext * context) {
if ((int) shape.size() > MAX_RANK)
throw std::invalid_argument("NDArrayFactory::create: rank of NDArray can't exceed 32 !");
ShapeDescriptor descriptor(nd4j::DataType::BOOL, order, shape);
if (descriptor.arrLength() != data.size()) {
nd4j_printf("NDArrayFactory::create: data size [%i] doesn't match shape length [%lld]\n", data.size(), descriptor.arrLength());
throw std::runtime_error("NDArrayFactory::create: data size doesn't match shape");
}
bool* hostBuffer = nullptr;
ALLOCATE(hostBuffer, context->getWorkspace(), data.size(), bool);
std::copy(data.begin(), data.end(), hostBuffer);
std::shared_ptr<DataBuffer> buffer = std::make_shared<DataBuffer>(hostBuffer, data.size() * sizeof(bool), nd4j::DataType::BOOL, true, context->getWorkspace());
NDArray result(buffer, descriptor, context);
return result;
}
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::vector<T> &data, nd4j::LaunchContext * context) {
if ((int) shape.size() > MAX_RANK)
throw std::invalid_argument("NDArrayFactory::create: rank of NDArray can't exceed 32 !");
ShapeDescriptor descriptor(DataTypeUtils::fromT<T>(), order, shape);
if (descriptor.arrLength() != data.size()) {
nd4j_printf("NDArrayFactory::create: data size [%i] doesn't match shape length [%lld]\n", data.size(), descriptor.arrLength());
throw std::runtime_error("NDArrayFactory::create: data size doesn't match shape");
}
std::shared_ptr<DataBuffer> buffer = std::make_shared<DataBuffer>(data.data(), DataTypeUtils::fromT<T>(), descriptor.arrLength() * sizeof(T), context->getWorkspace());
NDArray result(buffer, descriptor, context);
return result;
}
NDArray NDArrayFactory::string(const char *str, nd4j::LaunchContext * context) {
std::string s(str);
return string(s, context);
}
NDArray* NDArrayFactory::string_(const char *str, nd4j::LaunchContext * context) {
return string_(std::string(str), context);
}
NDArray NDArrayFactory::string(const std::string &str, nd4j::LaunchContext * context) {
auto headerLength = ShapeUtils::stringBufferHeaderRequirements(1);
std::shared_ptr<DataBuffer> pBuffer = std::make_shared<DataBuffer>(headerLength + str.length(), DataType::UTF8, context->getWorkspace(), true);
NDArray res(pBuffer, ShapeDescriptor::scalarDescriptor(DataType::UTF8), context);
int8_t* buffer = reinterpret_cast<int8_t*>(res.getBuffer());
auto offsets = reinterpret_cast<Nd4jLong *>(buffer);
offsets[0] = 0;
offsets[1] = str.length();
auto data = buffer + headerLength;
memcpy(data, str.c_str(), str.length());
res.tickWriteHost();
res.syncToDevice();
return res;
}
NDArray* NDArrayFactory::string_(const std::string &str, nd4j::LaunchContext * context) {
auto res = new NDArray();
*res = NDArrayFactory::string(str, context);
return res;
}
////////////////////////////////////////////////////////////////////////
template<typename T>
NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, nd4j::LaunchContext * context) {
return create_(order, shape, DataTypeUtils::fromT<T>(), context);
}
BUILD_SINGLE_TEMPLATE(template NDArray* NDArrayFactory::create_, (const char order, const std::vector<Nd4jLong> &shape, nd4j::LaunchContext * context), LIBND4J_TYPES);
////////////////////////////////////////////////////////////////////////
template <typename T>
void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<T> &vector) {
memcpy(ptr, vector.data(), vector.size() * sizeof(T));
}
template <>
void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<bool> &vector) {
auto p = reinterpret_cast<bool *>(ptr);
for (Nd4jLong e = 0; e < vector.size(); e++)
p[e] = vector[e];
}
template void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<double> &vector);
template void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<float> &vector);
template void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<float16> &vector);
template void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<Nd4jLong> &vector);
template void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<int> &vector);
template void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<int16_t> &vector);
template void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<uint8_t> &vector);
template void NDArrayFactory::memcpyFromVector(void *ptr, const std::vector<int8_t> &vector);
#ifndef __JAVACPP_HACK__
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const T value, const char order, nd4j::LaunchContext * context) {
return valueOf(std::vector<Nd4jLong>(shape), value, order);
}
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const double value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const float value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const float16 value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const bfloat16 value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const Nd4jLong value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const int value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const uint8_t value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const int8_t value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const int16_t value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::initializer_list<Nd4jLong>& shape, const bool value, const char order, nd4j::LaunchContext * context);
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<T>& data, nd4j::LaunchContext * context) {
std::vector<T> vec(data);
return create<T>(order, shape, vec, context);
}
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<double>& data, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<float>& data, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<float16>& data, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<bfloat16>& data, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<Nd4jLong>& data, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<int>& data, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<int16_t>& data, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<int8_t>& data, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<uint8_t>& data, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<bool>& data, nd4j::LaunchContext * context);
#endif
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray* NDArrayFactory::create_(const T scalar, nd4j::LaunchContext * context) {
std::shared_ptr<DataBuffer> buffer = std::make_shared<DataBuffer>(1 * sizeof(T), DataTypeUtils::fromT<T>(), context->getWorkspace(), true);
NDArray* res = new NDArray(buffer, ShapeDescriptor::scalarDescriptor(DataTypeUtils::fromT<T>()), context);
res->bufferAsT<T>()[0] = scalar;
res->tickWriteHost();
res->syncToDevice();
return res;
}
template NDArray* NDArrayFactory::create_(const double scalar, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const float scalar, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const float16 scalar, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const bfloat16 scalar, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const Nd4jLong scalar, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const int scalar, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const bool scalar, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const int8_t scalar, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const uint8_t scalar, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const int16_t scalar, nd4j::LaunchContext * context);
template <typename T>
NDArray NDArrayFactory::create(nd4j::DataType type, const T scalar, nd4j::LaunchContext * context) {
if (type == DataTypeUtils::fromT<T>())
return NDArrayFactory::create(scalar, context);
NDArray res(type, context);
res.p(0, scalar);
res.syncToDevice();
return res;
}
// BUILD_DOUBLE_TEMPLATE(template NDArray NDArrayFactory::create, (DataType type, const T scalar, nd4j::LaunchContext * context), LIBND4J_TYPES);
template NDArray NDArrayFactory::create(DataType type, const double scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(DataType type, const float scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(DataType type, const float16 scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(DataType type, const bfloat16 scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(DataType type, const Nd4jLong scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(DataType type, const int scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(DataType type, const int8_t scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(DataType type, const uint8_t scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(DataType type, const uint16_t scalar, nd4j::LaunchContext* workspace);
template NDArray NDArrayFactory::create(DataType type, const uint32_t scalar, nd4j::LaunchContext* workspace);
template NDArray NDArrayFactory::create(DataType type, const uint64_t scalar, nd4j::LaunchContext* workspace);
template NDArray NDArrayFactory::create(DataType type, const int16_t scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(DataType type, const bool scalar, nd4j::LaunchContext * context);
template <typename T>
NDArray NDArrayFactory::create(const T scalar, nd4j::LaunchContext * context) {
std::shared_ptr<DataBuffer> buffer = std::make_shared<DataBuffer>(1 * sizeof(T), DataTypeUtils::fromT<T>(), context->getWorkspace(), true);
NDArray res(buffer, ShapeDescriptor::scalarDescriptor(DataTypeUtils::fromT<T>()), context);
res.bufferAsT<T>()[0] = scalar;
res.tickWriteHost();
res.syncToDevice();
return res;
}
template NDArray NDArrayFactory::create(const double scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const float scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const float16 scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const bfloat16 scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const Nd4jLong scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const int scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const int8_t scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const uint8_t scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const int16_t scalar, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const uint16_t scalar, nd4j::LaunchContext* workspace);
template NDArray NDArrayFactory::create(const uint32_t scalar, nd4j::LaunchContext* workspace);
template NDArray NDArrayFactory::create(const uint64_t scalar, nd4j::LaunchContext* workspace);
template NDArray NDArrayFactory::create(const bool scalar, nd4j::LaunchContext * context);
////////////////////////////////////////////////////////////////////////
template<typename T>
NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<T> &data, nd4j::LaunchContext * context) {
return new NDArray(NDArrayFactory::create<T>(order, shape, data, context));
}
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<double> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<float> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<float16> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<bfloat16> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<int> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<unsigned int> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<unsigned long> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<Nd4jLong> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<int8_t> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<uint8_t> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<int16_t> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<uint16_t> &data, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::create_(const char order, const std::vector<Nd4jLong> &shape, const std::vector<bool> &data, nd4j::LaunchContext * context);
////////////////////////////////////////////////////////////////////////
template <>
NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, NDArray* value, const char order, nd4j::LaunchContext * context) {
auto result = create_(order, shape, value->dataType(), context);
result->assign(*value);
return result;
}
template <>
NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, NDArray& value, const char order, nd4j::LaunchContext * context) {
auto result = create_(order, shape, value.dataType(), context);
result->assign(value);
return result;
}
template <typename T>
NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const T value, const char order, nd4j::LaunchContext * context) {
auto result = create_(order, shape, DataTypeUtils::fromT<T>());
result->assign(value);
return result;
}
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const double value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const float value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const float16 value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const bfloat16 value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const Nd4jLong value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const int value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const int16_t value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const int8_t value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const uint8_t value, const char order, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const bool value, const char order, nd4j::LaunchContext * context);
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray* NDArrayFactory::linspace(const T from, const T to, const Nd4jLong numElements) {
NDArray* result = NDArrayFactory::vector<T>(numElements);
//TO DO: linspace should be executed on DEVICE, but only CPU version implemnted!
for (Nd4jLong e = 0; e < numElements; e++) {
T step = (T) e / ((T) numElements - (T) 1);
result->p<T >(e, (from * ((T) 1 - step) + step * to));
}
result->syncToDevice();
return result;
}
template NDArray* NDArrayFactory::linspace(const double from, const double to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const float from, const float to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const float16 from, const float16 to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const bfloat16 from, const bfloat16 to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const Nd4jLong from, const Nd4jLong to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const int from, const int to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const int16_t from, const int16_t to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const uint8_t from, const uint8_t to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const uint16_t from, const uint16_t to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const uint32_t from, const uint32_t to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const uint64_t from, const uint64_t to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const int8_t from, const int8_t to, const Nd4jLong numElements);
template NDArray* NDArrayFactory::linspace(const bool from, const bool to, const Nd4jLong numElements);
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray* NDArrayFactory::vector(Nd4jLong length, const T value, nd4j::LaunchContext * context) {
std::shared_ptr<DataBuffer> buffer = std::make_shared<DataBuffer>(length * sizeof(T), DataTypeUtils::fromT<T>(), context->getWorkspace(), true);
auto res = new NDArray(buffer, ShapeDescriptor::vectorDescriptor(length, DataTypeUtils::fromT<T>()), context);
if (value == (T)0.0f)
res->nullify();
else
res->assign(value);
return res;
}
template NDArray* NDArrayFactory::vector(Nd4jLong length, const double startingValue, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const float startingValue, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const float16 startingValue, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const bfloat16 startingValue, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const Nd4jLong startingValue, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const int startingValue, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const uint8_t startingValue, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const uint16_t startingValue, nd4j::LaunchContext *workspace);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const uint32_t startingValue, nd4j::LaunchContext *workspace);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const uint64_t startingValue, nd4j::LaunchContext *workspace);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const int8_t startingValue, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const int16_t startingValue, nd4j::LaunchContext * context);
template NDArray* NDArrayFactory::vector(Nd4jLong length, const bool startingValue, nd4j::LaunchContext * context);
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray NDArrayFactory::create(const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context) {
std::vector<Nd4jLong> vec(shape);
return create<T>(order, vec, context);
}
BUILD_SINGLE_TEMPLATE(template NDArray NDArrayFactory::create, (const char, const std::initializer_list<Nd4jLong>&, nd4j::LaunchContext * context), LIBND4J_TYPES);
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, nd4j::LaunchContext * context) {
return create(order, shape, DataTypeUtils::fromT<T>(), context);
}
BUILD_SINGLE_TEMPLATE(template NDArray NDArrayFactory::create, (const char order, const std::vector<Nd4jLong> &shape, nd4j::LaunchContext * context), LIBND4J_TYPES);
////////////////////////////////////////////////////////////////////////
NDArray NDArrayFactory::create(const char order, const std::vector<Nd4jLong> &shape, nd4j::DataType dtype, nd4j::LaunchContext* context) {
if ((int) shape.size() > MAX_RANK)
throw std::invalid_argument("NDArrayFactory::create: rank of NDArray can't exceed 32");
ShapeDescriptor descriptor(dtype, order, shape);
std::shared_ptr<DataBuffer> buffer = std::make_shared<DataBuffer>(descriptor.arrLength() * DataTypeUtils::sizeOfElement(dtype), dtype, context->getWorkspace());
NDArray result(buffer, descriptor, context);
result.nullify();
return result;
}
////////////////////////////////////////////////////////////////////////
NDArray NDArrayFactory::create(nd4j::DataType dtype, nd4j::LaunchContext * context) {
std::shared_ptr<DataBuffer> buffer = std::make_shared<DataBuffer>(DataTypeUtils::sizeOfElement(dtype), dtype, context->getWorkspace(), true);
NDArray res(buffer, ShapeDescriptor::scalarDescriptor(dtype), context);
res.nullify();
return res;
}
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray NDArrayFactory::create(const std::vector<T> &values, nd4j::LaunchContext * context) {
std::shared_ptr<DataBuffer> buffer = std::make_shared<DataBuffer>(values.size() * sizeof(T), DataTypeUtils::fromT<T>(), context->getWorkspace(), true);
NDArray res(buffer, ShapeDescriptor::vectorDescriptor(values.size(), DataTypeUtils::fromT<T>()), context);
memcpyFromVector<T>(res.getBuffer(), values);
res.tickWriteHost();
res.syncToDevice();
return res;
}
template NDArray NDArrayFactory::create(const std::vector<double> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<float> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<float16> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<bfloat16> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<Nd4jLong> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<int> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<int16_t> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<uint16_t> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<int8_t> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<uint8_t> &values, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(const std::vector<bool> &values, nd4j::LaunchContext * context);
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray* NDArrayFactory::empty_(nd4j::LaunchContext * context) {
auto shapeInfo = ShapeBuilders::createScalarShapeInfo(DataTypeUtils::fromT<T>(), context->getWorkspace());
ArrayOptions::setPropertyBit(shapeInfo, ARRAY_EMPTY);
auto result = new NDArray(nullptr, shapeInfo, context, false);
RELEASE(shapeInfo, context->getWorkspace());
return result;
}
BUILD_SINGLE_TEMPLATE(template NDArray* NDArrayFactory::empty_, (nd4j::LaunchContext * context), LIBND4J_TYPES);
NDArray* NDArrayFactory::empty_(nd4j::DataType dataType, nd4j::LaunchContext * context) {
if (context == nullptr)
context = nd4j::LaunchContext ::defaultContext();
auto shapeInfo = ShapeBuilders::createScalarShapeInfo(dataType, context->getWorkspace());
ArrayOptions::setPropertyBit(shapeInfo, ARRAY_EMPTY);
auto result = new NDArray(nullptr, shapeInfo, context, false);
RELEASE(shapeInfo, context->getWorkspace());
return result;
}
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray NDArrayFactory::empty(nd4j::LaunchContext * context) {
return empty(DataTypeUtils::fromT<T>(), context);
}
BUILD_SINGLE_TEMPLATE(template NDArray NDArrayFactory::empty, (nd4j::LaunchContext * context), LIBND4J_TYPES);
////////////////////////////////////////////////////////////////////////
NDArray NDArrayFactory::empty(nd4j::DataType dataType, nd4j::LaunchContext * context) {
auto shapeInfo = ShapeBuilders::createScalarShapeInfo(dataType, context->getWorkspace());
ArrayOptions::setPropertyBit(shapeInfo, ARRAY_EMPTY);
NDArray result(nullptr, shapeInfo, context, false);
RELEASE(shapeInfo, context->getWorkspace());
return result;
}
////////////////////////////////////////////////////////////////////////
NDArray* NDArrayFactory::valueOf(const std::vector<Nd4jLong>& shape, const NDArray& value, const char order, nd4j::LaunchContext * context) {
auto res = NDArrayFactory::create_(order, shape, value.dataType(), context);
res->assign(const_cast<NDArray&>(value));
return res;
}
////////////////////////////////////////////////////////////////////////
NDArray* NDArrayFactory::create_( const char order, const std::vector<Nd4jLong> &shape, nd4j::DataType dataType, nd4j::LaunchContext * context) {
return new NDArray(order, shape, dataType, context);
}
////////////////////////////////////////////////////////////////////////
template <typename T>
NDArray NDArrayFactory::create(T* buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context) {
if ((int) shape.size() > MAX_RANK)
throw std::invalid_argument("NDArrayFactory::create: Rank of NDArray can't exceed 32");
std::vector<Nd4jLong> shp(shape);
ShapeDescriptor descriptor(DataTypeUtils::fromT<T>(), order, shp);
std::shared_ptr<DataBuffer> pBuffer = std::make_shared<DataBuffer>(buffer, descriptor.arrLength() * sizeof(T), descriptor.dataType(), false, context->getWorkspace());
NDArray result(pBuffer, descriptor, context);
return result;
}
template NDArray NDArrayFactory::create(double* buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(float* buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(float16* buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(bfloat16* buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(Nd4jLong * buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(int* buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(bool* buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(uint8_t * buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(int8_t* buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
template NDArray NDArrayFactory::create(int16_t* buffer, const char order, const std::initializer_list<Nd4jLong>& shape, nd4j::LaunchContext * context);
NDArray NDArrayFactory::string(char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<const char *> &strings, nd4j::LaunchContext * context) {
std::vector<const char*> vec(strings);
return NDArrayFactory::string(order, shape, vec, context);
}
NDArray NDArrayFactory::string(char order, const std::vector<Nd4jLong> &shape, const std::vector<const char *> &strings, nd4j::LaunchContext * context) {
std::vector<std::string> vec(strings.size());
int cnt = 0;
for (auto s:strings)
vec[cnt++] = std::string(s);
return NDArrayFactory::string(order, shape, vec, context);
}
NDArray NDArrayFactory::string(char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<std::string> &string, nd4j::LaunchContext * context) {
std::vector<std::string> vec(string);
return NDArrayFactory::string(order, shape, vec, context);
}
NDArray* NDArrayFactory::string_(char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<const char *> &strings, nd4j::LaunchContext * context) {
std::vector<const char*> vec(strings);
return NDArrayFactory::string_(order, shape, vec, context);
}
NDArray* NDArrayFactory::string_(char order, const std::vector<Nd4jLong> &shape, const std::vector<const char *> &strings, nd4j::LaunchContext * context) {
std::vector<std::string> vec(strings.size());
int cnt = 0;
for (auto s:strings)
vec[cnt++] = std::string(s);
return NDArrayFactory::string_(order, shape, vec, context);
}
NDArray* NDArrayFactory::string_(char order, const std::vector<Nd4jLong> &shape, const std::initializer_list<std::string> &string, nd4j::LaunchContext * context) {
std::vector<std::string> vec(string);
return NDArrayFactory::string_(order, shape, vec, context);
}
NDArray NDArrayFactory::string(char order, const std::vector<Nd4jLong> &shape, const std::vector<std::string> &string, nd4j::LaunchContext * context) {
if (context == nullptr)
context = nd4j::LaunchContext ::defaultContext();
auto headerLength = ShapeUtils::stringBufferHeaderRequirements(string.size());
std::vector<Nd4jLong> offsets(string.size() + 1);
Nd4jLong dataLength = 0;
for (int e = 0; e < string.size(); e++) {
offsets[e] = dataLength;
dataLength += string[e].length();
}
offsets[string.size()] = dataLength;
std::shared_ptr<DataBuffer> pBuffer = std::make_shared<DataBuffer>(headerLength + dataLength, DataType::UTF8, context->getWorkspace(), true);
NDArray res(pBuffer, ShapeDescriptor(DataType::UTF8, order, shape), context);
res.setAttached(context->getWorkspace() != nullptr);
if (res.lengthOf() != string.size())
throw std::invalid_argument("Number of strings should match length of array");
memcpy(res.buffer(), offsets.data(), offsets.size() * sizeof(Nd4jLong));
auto data = static_cast<int8_t*>(res.buffer()) + headerLength;
int resLen = res.lengthOf();
for (int e = 0; e < resLen; e++) {
auto length = offsets[e+1] - offsets[e];
auto cdata = data + offsets[e];
memcpy(cdata, string[e].c_str(), string[e].length());
}
res.tickWriteHost();
res.syncToDevice();
return res;
}
NDArray* NDArrayFactory::string_(char order, const std::vector<Nd4jLong> &shape, const std::vector<std::string> &string, nd4j::LaunchContext * context) {
auto res = new NDArray();
*res = NDArrayFactory::string(order, shape, string, context);
return res;
}
}