Logo ROOT  
Reference Guide
 
Loading...
Searching...
No Matches
CpuBuffer.cxx
Go to the documentation of this file.
1// @(#)root/tmva/tmva/dnn:$Id$
2// Author: Simon Pfreundschuh 12/08/16
3
4/*************************************************************************
5 * Copyright (C) 2016, Simon Pfreundschuh *
6 * All rights reserved. *
7 * *
8 * For the licensing terms see $ROOTSYS/LICENSE. *
9 * For the list of contributors see $ROOTSYS/README/CREDITS. *
10 *************************************************************************/
11
12/////////////////////////////////////////////////////////////
13// CPU Buffer interface class for the generic data loader. //
14/////////////////////////////////////////////////////////////
15
16#include <vector>
17#include <memory>
18#include "TMVA/DataSetInfo.h"
19#include "TMVA/DNN/DataLoader.h"
22#include "Rtypes.h"
23#include <iostream>
24
25namespace TMVA {
26namespace DNN {
27
28//______________________________________________________________________________
29template <typename AReal>
31{
32 delete[] * pointer;
33 delete[] pointer;
34}
35
36//______________________________________________________________________________
37template <typename AReal>
39{
40 AReal **pointer = new AReal *[1];
41 *pointer = new AReal[size];
42 fBuffer = std::shared_ptr<AReal *>(pointer, fDestructor);
43}
44
45//______________________________________________________________________________
46template <typename AReal>
48{
49 TCpuBuffer buffer = *this;
50 buffer.fOffset = offset;
51 buffer.fSize = size;
52 return buffer;
53}
54
55//______________________________________________________________________________
56template <typename AReal>
58{
59 //std::copy*this->fBuffer, *other.fBuffer);
60 std::copy( *other.fBuffer, *other.fBuffer+fSize, *this->fBuffer);
61}
62
63//______________________________________________________________________________
64template <typename AReal>
66{
67 std::copy( *this->fBuffer, *this->fBuffer+fSize, *other.fBuffer);
68 //std::swap(*this->fBuffer, *other.fBuffer);
69}
70
71//______________________________________________________________________________
72template <>
74 size_t batchSize)
75{
76 const TMatrixT<Float_t> &inputMatrix = std::get<0>(fData);
77 size_t n = inputMatrix.GetNcols();
78
79 for (size_t i = 0; i < batchSize; i++) {
80 size_t sampleIndex = *sampleIterator;
81 for (size_t j = 0; j < n; j++) {
82 size_t bufferIndex = j * batchSize + i;
83 buffer[bufferIndex] = static_cast<Float_t>(inputMatrix(sampleIndex, j));
84 }
85 sampleIterator++;
86 }
87}
88
89//______________________________________________________________________________
90template <>
92 size_t batchSize)
93{
94 const TMatrixT<Float_t> &outputMatrix = std::get<1>(fData);
95 size_t n = outputMatrix.GetNcols();
96
97 for (size_t i = 0; i < batchSize; i++) {
98 size_t sampleIndex = *sampleIterator;
99 for (size_t j = 0; j < n; j++) {
100 size_t bufferIndex = j * batchSize + i;
101 buffer[bufferIndex] = static_cast<Float_t>(outputMatrix(sampleIndex, j));
102 }
103 sampleIterator++;
104 }
105}
106
107//______________________________________________________________________________
108template <>
110 size_t batchSize)
111{
112 const TMatrixT<Float_t> &outputMatrix = std::get<2>(fData);
113
114 for (size_t i = 0; i < batchSize; i++) {
115 size_t sampleIndex = *sampleIterator;
116 buffer[i] = static_cast<Float_t>(outputMatrix(sampleIndex, 0));
117 sampleIterator++;
118 }
119}
120
121//______________________________________________________________________________
122template <>
124 size_t batchSize)
125{
126 const TMatrixT<Double_t> &inputMatrix = std::get<0>(fData);
127 size_t n = inputMatrix.GetNcols();
128
129 for (size_t i = 0; i < batchSize; i++) {
130 size_t sampleIndex = *sampleIterator;
131 for (size_t j = 0; j < n; j++) {
132 size_t bufferIndex = j * batchSize + i;
133 buffer[bufferIndex] = inputMatrix(sampleIndex, j);
134 }
135 sampleIterator++;
136 }
137}
138
139//______________________________________________________________________________
140template <>
142 IndexIterator_t sampleIterator, size_t batchSize)
143{
144 const TMatrixT<Double_t> &outputMatrix = std::get<1>(fData);
145 size_t n = outputMatrix.GetNcols();
146
147 for (size_t i = 0; i < batchSize; i++) {
148 size_t sampleIndex = *sampleIterator;
149 for (size_t j = 0; j < n; j++) {
150 size_t bufferIndex = j * batchSize + i;
151 buffer[bufferIndex] = outputMatrix(sampleIndex, j);
152 }
153 sampleIterator++;
154 }
155}
156
157//______________________________________________________________________________
158template <>
160 IndexIterator_t sampleIterator, size_t batchSize)
161{
162 const TMatrixT<Double_t> &outputMatrix = std::get<2>(fData);
163
164 for (size_t i = 0; i < batchSize; i++) {
165 size_t sampleIndex = *sampleIterator;
166 buffer[i] = static_cast<Double_t>(outputMatrix(sampleIndex, 0));
167 sampleIterator++;
168 }
169}
170
171//______________________________________________________________________________
172template <>
174 size_t batchSize)
175{
176 Event *event = std::get<0>(fData)[0];
177 size_t n = event->GetNVariables();
178 for (size_t i = 0; i < batchSize; i++) {
179 size_t sampleIndex = * sampleIterator++;
180 event = std::get<0>(fData)[sampleIndex];
181 for (size_t j = 0; j < n; j++) {
182 size_t bufferIndex = j * batchSize + i;
183 buffer[bufferIndex] = event->GetValue(j);
184 }
185 }
186}
187
188//______________________________________________________________________________
189template <>
191 size_t batchSize)
192{
193 const DataSetInfo &info = std::get<1>(fData);
194 size_t n = buffer.GetSize() / batchSize;
195
196 // Copy target(s).
197
198 for (size_t i = 0; i < batchSize; i++) {
199 size_t sampleIndex = *sampleIterator++;
200 Event *event = std::get<0>(fData)[sampleIndex];
201 for (size_t j = 0; j < n; j++) {
202 // Copy output matrices.
203 size_t bufferIndex = j * batchSize + i;
204 // Classification
205 if (event->GetNTargets() == 0) {
206 if (n == 1) {
207 // Binary.
208 buffer[bufferIndex] = (info.IsSignal(event)) ? 1.0 : 0.0;
209 } else {
210 // Multiclass.
211 buffer[bufferIndex] = 0.0;
212 if (j == event->GetClass()) {
213 buffer[bufferIndex] = 1.0;
214 }
215 }
216 } else {
217 buffer[bufferIndex] = static_cast<Float_t>(event->GetTarget(j));
218 }
219 }
220 }
221}
222
223//______________________________________________________________________________
224template <>
226 size_t batchSize)
227{
228 for (size_t i = 0; i < batchSize; i++) {
229 size_t sampleIndex = *sampleIterator++;
230 Event *event = std::get<0>(fData)[sampleIndex];
231 buffer[i] = event->GetWeight();
232 }
233}
234
235//______________________________________________________________________________
236template <>
238 size_t batchSize)
239{
240 Event *event = std::get<0>(fData)[0];
241 size_t n = event->GetNVariables();
242 for (size_t i = 0; i < batchSize; i++) {
243 size_t sampleIndex = * sampleIterator++;
244 event = std::get<0>(fData)[sampleIndex];
245 for (size_t j = 0; j < n; j++) {
246 size_t bufferIndex = j * batchSize + i;
247 buffer[bufferIndex] = static_cast<Float_t>(event->GetValue(j));
248 }
249 }
250}
251
252//______________________________________________________________________________
253template <>
255 size_t batchSize)
256{
257 const DataSetInfo &info = std::get<1>(fData);
258 size_t n = buffer.GetSize() / batchSize;
259
260 // Copy target(s).
261
262 for (size_t i = 0; i < batchSize; i++) {
263 size_t sampleIndex = *sampleIterator++;
264 Event *event = std::get<0>(fData)[sampleIndex];
265 for (size_t j = 0; j < n; j++) {
266 // Copy output matrices.
267 size_t bufferIndex = j * batchSize + i;
268 // Classification
269 if (event->GetNTargets() == 0) {
270 if (n == 1) {
271 // Binary.
272 buffer[bufferIndex] = (info.IsSignal(event)) ? 1.0 : 0.0;
273 } else {
274 // Multiclass.
275 buffer[bufferIndex] = 0.0;
276 if (j == event->GetClass()) {
277 buffer[bufferIndex] = 1.0;
278 }
279 }
280 } else {
281 buffer[bufferIndex] = static_cast<Float_t>(event->GetTarget(j));
282 }
283 }
284 }
285}
286
287//______________________________________________________________________________
288template <>
290 size_t batchSize)
291{
292 for (size_t i = 0; i < batchSize; i++) {
293 size_t sampleIndex = *sampleIterator++;
294 Event *event = std::get<0>(fData)[sampleIndex];
295 buffer[i] = static_cast<Float_t>(event->GetWeight());
296 }
297}
298
299//______________________________________________________________________________
300template <>
302 IndexIterator_t sampleIterator)
303{
304 const std::vector<TMatrixT<Double_t>> &inputTensor = std::get<0>(fData);
305
306 if (fBatchDepth == 1) {
307 for (size_t i = 0; i < fBatchHeight; i++) {
308 size_t sampleIndex = *sampleIterator;
309 for (size_t j = 0; j < fBatchWidth; j++) {
310 size_t bufferIndex = j * fBatchHeight + i;
311 buffer[bufferIndex] = static_cast<Float_t>(inputTensor[0](sampleIndex, j));
312 }
313 sampleIterator++;
314 }
315 } else {
316 for (size_t i = 0; i < fBatchDepth; i++) {
317 size_t sampleIndex = *sampleIterator;
318 for (size_t j = 0; j < fBatchHeight; j++) {
319 for (size_t k = 0; k < fBatchWidth; k++) {
320 size_t bufferIndex = i * fBatchHeight * fBatchWidth + k * fBatchHeight + j;
321 buffer[bufferIndex] = static_cast<Float_t>(inputTensor[sampleIndex](j, k));
322 }
323 }
324 sampleIterator++;
325 }
326 }
327}
328
329//______________________________________________________________________________
330template <>
332 IndexIterator_t sampleIterator)
333{
334 const TMatrixT<Double_t> &outputMatrix = std::get<1>(fData);
335 size_t n = outputMatrix.GetNcols();
336
337 for (size_t i = 0; i < fBatchSize; i++) {
338 size_t sampleIndex = *sampleIterator;
339 for (size_t j = 0; j < n; j++) {
340 size_t bufferIndex = j * fBatchSize + i;
341 buffer[bufferIndex] = static_cast<Float_t>(outputMatrix(sampleIndex, j));
342 }
343 sampleIterator++;
344 }
345}
346
347//______________________________________________________________________________
348template <>
350 IndexIterator_t sampleIterator)
351{
352 const TMatrixT<Double_t> &outputMatrix = std::get<2>(fData);
353
354 for (size_t i = 0; i < fBatchSize; i++) {
355 size_t sampleIndex = *sampleIterator;
356 buffer[i] = static_cast<Float_t>(outputMatrix(sampleIndex, 0));
357 sampleIterator++;
358 }
359}
360
361#if 0
362//______________________________________________________________________________
363template <>
365{
366 // After copying the data to the device, wrap the device buffer in the respective
367 // architectures matrix type
368 DeviceBufferTuple DeviceBuffers = CopyTensorBatches();
369
370 Tensor_t inputTensor( std::get<0>(DeviceBuffers), { fBatchHeight, fBatchWidth, fBatchSize } );
371 // size_t jump = fBatchHeight * fBatchWidth;
372 // for (size_t i = 0; i < fBatchSize; i++) {
373 // DeviceBuffer_t subInputDeviceBuffer = std::get<0>(DeviceBuffers).GetSubBuffer(i * jump, jump);
374 // inputTensor.emplace_back(subInputDeviceBuffer, fBatchHeight, fBatchWidth);
375 // }
376 Matrix_t outputMatrix(std::get<1>(DeviceBuffers), fBatchSize, fNOutputFeatures);
377 Matrix_t weightMatrix(std::get<2>(DeviceBuffers), fBatchSize, fNOutputFeatures);
378
379 fBatchIndex++;
380 return TTensorBatch<TCpu<Float_t> >(inputTensor, outputMatrix, weightMatrix);
381}
382#endif
383
384//______________________________________________________________________________
385template <>
387 IndexIterator_t sampleIterator)
388{
389 const std::vector<TMatrixT<Double_t>> &inputTensor = std::get<0>(fData);
390
391 if (fBatchDepth == 1) {
392 for (size_t i = 0; i < fBatchHeight; i++) {
393 size_t sampleIndex = *sampleIterator;
394 for (size_t j = 0; j < fBatchWidth; j++) {
395 size_t bufferIndex = j * fBatchHeight + i;
396 buffer[bufferIndex] = inputTensor[0](sampleIndex, j);
397 }
398 sampleIterator++;
399 }
400 } else {
401 for (size_t i = 0; i < fBatchDepth; i++) {
402 size_t sampleIndex = *sampleIterator;
403 for (size_t j = 0; j < fBatchHeight; j++) {
404 for (size_t k = 0; k < fBatchWidth; k++) {
405 size_t bufferIndex = i * fBatchHeight * fBatchWidth + k * fBatchHeight + j;
406 buffer[bufferIndex] = inputTensor[sampleIndex](j, k);
407 }
408 }
409 sampleIterator++;
410 }
411 }
412}
413
414//______________________________________________________________________________
415template <>
417 IndexIterator_t sampleIterator)
418{
419 const TMatrixT<Double_t> &outputMatrix = std::get<1>(fData);
420 size_t n = outputMatrix.GetNcols();
421
422 for (size_t i = 0; i < fBatchSize; i++) {
423 size_t sampleIndex = *sampleIterator;
424 for (size_t j = 0; j < n; j++) {
425 size_t bufferIndex = j * fBatchSize + i;
426 buffer[bufferIndex] = outputMatrix(sampleIndex, j);
427 }
428 sampleIterator++;
429 }
430}
431
432//______________________________________________________________________________
433template <>
435 IndexIterator_t sampleIterator)
436{
437 const TMatrixT<Double_t> &outputMatrix = std::get<2>(fData);
438
439 for (size_t i = 0; i < fBatchSize; i++) {
440 size_t sampleIndex = *sampleIterator;
441 buffer[i] = static_cast<Double_t>(outputMatrix(sampleIndex, 0));
442 sampleIterator++;
443 }
444}
445#if 0
446//______________________________________________________________________________
447template <>
449{
450 // After copying the data to the device, wrap the device buffer in the respective
451 // architectures matrix type
452 DeviceBufferTuple DeviceBuffers = CopyTensorBatches();
453
454 Tensor_t inputTensor( std::get<0>(DeviceBuffers), { fBatchHeight, fBatchWidth, fBatchSize } );
455 // size_t jump = fBatchHeight * fBatchWidth;
456 // for (size_t i = 0; i < fBatchSize; i++) {
457 // DeviceBuffer_t subInputDeviceBuffer = std::get<0>(DeviceBuffers).GetSubBuffer(i * jump, jump);
458 // inputTensor.emplace_back(subInputDeviceBuffer, fBatchHeight, fBatchWidth);
459 // }
460
461 Matrix_t outputMatrix(std::get<1>(DeviceBuffers), fBatchSize, fNOutputFeatures);
462 Matrix_t weightMatrix(std::get<2>(DeviceBuffers), fBatchSize, fNOutputFeatures);
463
464 fBatchIndex++;
465 return TTensorBatch<TCpu<Double_t> >(inputTensor, outputMatrix, weightMatrix);
466}
467#endif
468
469
470///- re-implement specialization for Double_t
471//______________________________________________________________________________
472template <>
474 IndexIterator_t sampleIterator)
475{
476 // one event, one example in the batch
477
478 if (fBatchDepth == 1 && fBatchHeight == fBatchSize) {
479 for (size_t i = 0; i < fBatchHeight; i++) {
480 size_t sampleIndex = *sampleIterator;
481 Event * event = std::get<0>(fData)[sampleIndex];
482 for (size_t j = 0; j < fBatchWidth; j++) {
483 size_t bufferIndex = j * fBatchHeight + i;
484 buffer[bufferIndex] = event->GetValue(j);
485 }
486 sampleIterator++;
487 }
488 } else if (fBatchDepth == fBatchSize) {
489 // batchDepth is batch size
490 for (size_t i = 0; i < fBatchDepth; i++) {
491 size_t sampleIndex = *sampleIterator;
492 Event * event = std::get<0>(fData)[sampleIndex];
493 for (size_t j = 0; j < fBatchHeight; j++) {
494 for (size_t k = 0; k < fBatchWidth; k++) {
495 // because of the ordering of tensor in memory is NHWC
496 size_t bufferIndex = i * fBatchHeight * fBatchWidth + k * fBatchHeight + j;
497 buffer[bufferIndex] = event->GetValue(j * fBatchWidth + k);
498 }
499 }
500 sampleIterator++;
501 }
502 }
503 else {
504 Error("TTensorDataLoader","Inconsistency between batch depth and batch size");
505 R__ASSERT(0); // one event, one example in the batch
506 }
507}
508
509//______________________________________________________________________________
510template <>
512 IndexIterator_t sampleIterator)
513{
514 const DataSetInfo &info = std::get<1>(fData);
515 size_t n = buffer.GetSize() / fBatchSize;
516
517 // Copy target(s).
518
519 for (size_t i = 0; i < fBatchSize; i++) {
520 size_t sampleIndex = *sampleIterator++;
521 Event *event = std::get<0>(fData)[sampleIndex];
522 for (size_t j = 0; j < n; j++) {
523 // Copy output matrices.
524 size_t bufferIndex = j * fBatchSize + i;
525 // Classification
526 if (event->GetNTargets() == 0) {
527 if (n == 1) {
528 // Binary.
529 buffer[bufferIndex] = (info.IsSignal(event)) ? 1.0 : 0.0;
530 } else {
531 // Multiclass.
532 buffer[bufferIndex] = 0.0;
533 if (j == event->GetClass()) {
534 buffer[bufferIndex] = 1.0;
535 }
536 }
537 } else {
538 buffer[bufferIndex] = static_cast<Float_t>(event->GetTarget(j));
539 }
540 }
541 }
542}
543
544//______________________________________________________________________________
545template <>
547 IndexIterator_t sampleIterator)
548{
549 for (size_t i = 0; i < fBatchSize; i++) {
550 size_t sampleIndex = *sampleIterator++;
551 Event *event = std::get<0>(fData)[sampleIndex];
552 buffer[i] = event->GetWeight();
553 }
554}
555
556#if 0
557//______________________________________________________________________________
558template <>
560{
561 // After copying the data to the device, wrap the device buffer in the respective
562 // architectures matrix type
563 DeviceBufferTuple DeviceBuffers = CopyTensorBatches();
564
565
566 Tensor_t inputTensor( std::get<0>(DeviceBuffers), { fBatchHeight, fBatchWidth, fBatchSize } );
567 // size_t jump = fBatchHeight * fBatchWidth;
568 // for (size_t i = 0; i < fBatchSize; i++) {
569 // DeviceBuffer_t subInputDeviceBuffer = std::get<0>(DeviceBuffers).GetSubBuffer(i * jump, jump);
570 // inputTensor.emplace_back(subInputDeviceBuffer, fBatchHeight, fBatchWidth);
571 // }
572 Matrix_t outputMatrix(std::get<1>(DeviceBuffers), fBatchSize, fNOutputFeatures);
573 Matrix_t weightMatrix(std::get<2>(DeviceBuffers), fBatchSize, fNOutputFeatures);
574
575 fBatchIndex++;
576 return TTensorBatch<TCpu<Double_t> >(inputTensor, outputMatrix, weightMatrix);
577}
578#endif
579
580///- re-implement specialization for Float_t
581//______________________________________________________________________________
582template <>
584 IndexIterator_t sampleIterator)
585{
586 // one event, one example in the batch
587
588 if (fBatchDepth == 1 && fBatchHeight == fBatchSize) {
589 for (size_t i = 0; i < fBatchHeight; i++) {
590 size_t sampleIndex = *sampleIterator;
591 Event * event = std::get<0>(fData)[sampleIndex];
592 for (size_t j = 0; j < fBatchWidth; j++) {
593 size_t bufferIndex = j * fBatchHeight + i;
594 buffer[bufferIndex] = event->GetValue(j);
595 }
596 sampleIterator++;
597 }
598 } else if (fBatchDepth == fBatchSize) {
599 // batchDepth is batch size
600 for (size_t i = 0; i < fBatchDepth; i++) {
601 size_t sampleIndex = *sampleIterator;
602 Event * event = std::get<0>(fData)[sampleIndex];
603 for (size_t j = 0; j < fBatchHeight; j++) {
604 for (size_t k = 0; k < fBatchWidth; k++) {
605 // because of the column-major ordering
606 size_t bufferIndex = i * fBatchHeight * fBatchWidth + k * fBatchHeight + j;
607 buffer[bufferIndex] = event->GetValue(j * fBatchWidth + k);
608 }
609 }
610 sampleIterator++;
611 }
612 }
613 else {
614 Error("TTensorDataLoader","Inconsistency between batch depth and batch size");
615 R__ASSERT(0);
616 }
617}
618
619//______________________________________________________________________________
620template <>
622 IndexIterator_t sampleIterator)
623{
624 const DataSetInfo &info = std::get<1>(fData);
625 size_t n = buffer.GetSize() / fBatchSize;
626
627 // Copy target(s).
628
629 for (size_t i = 0; i < fBatchSize; i++) {
630 size_t sampleIndex = *sampleIterator++;
631 Event *event = std::get<0>(fData)[sampleIndex];
632 for (size_t j = 0; j < n; j++) {
633 // Copy output matrices.
634 size_t bufferIndex = j * fBatchSize + i;
635 // Classification
636 if (event->GetNTargets() == 0) {
637 if (n == 1) {
638 // Binary.
639 buffer[bufferIndex] = (info.IsSignal(event)) ? 1.0 : 0.0;
640 } else {
641 // Multiclass.
642 buffer[bufferIndex] = 0.0;
643 if (j == event->GetClass()) {
644 buffer[bufferIndex] = 1.0;
645 }
646 }
647 } else {
648 buffer[bufferIndex] = static_cast<Float_t>(event->GetTarget(j));
649 }
650 }
651 }
652}
653
654//______________________________________________________________________________
655template <>
657 IndexIterator_t sampleIterator)
658{
659 for (size_t i = 0; i < fBatchSize; i++) {
660 size_t sampleIndex = *sampleIterator++;
661 Event *event = std::get<0>(fData)[sampleIndex];
662 buffer[i] = event->GetWeight();
663 }
664}
665
666#if 0
667//______________________________________________________________________________
668template <>
670{
671 // After copying the data to the device, wrap the device buffer in the respective
672 // architectures matrix type
673 DeviceBufferTuple DeviceBuffers = CopyTensorBatches();
674
675 Tensor_t inputTensor( std::get<0>(DeviceBuffers), { fBatchHeight, fBatchWidth, fBatchSize } );
676 // std::vector<Matrix_t> inputTensor;
677 // size_t jump = fBatchHeight * fBatchWidth;
678 // for (size_t i = 0; i < fBatchSize; i++) {
679 // DeviceBuffer_t subInputDeviceBuffer = std::get<0>(DeviceBuffers).GetSubBuffer(i * jump, jump);
680 // inputTensor.emplace_back(subInputDeviceBuffer, fBatchHeight, fBatchWidth);
681 // }
682 Matrix_t outputMatrix(std::get<1>(DeviceBuffers), fBatchSize, fNOutputFeatures);
683 Matrix_t weightMatrix(std::get<2>(DeviceBuffers), fBatchSize, fNOutputFeatures);
684
685 fBatchIndex++;
686 return TTensorBatch<TCpu<Float_t> >(inputTensor, outputMatrix, weightMatrix);
687}
688#endif
689
690//______________________________________________________________________________
691// Explicit instantiations.
692template class TCpuBuffer<Double_t>;
693template class TCpuBuffer<Float_t>;
694
695template class TTensorDataLoader<TensorInput, TCpu<Float_t>>;
696template class TTensorDataLoader<TMVAInput_t, TCpu<Float_t>>;
697template class TTensorDataLoader<TensorInput, TCpu<Double_t>>;
698template class TTensorDataLoader<TMVAInput_t, TCpu<Double_t>>;
699
700} // namespace DNN
701} // namespace TMVA
fBuffer
dim_t fSize
size_t size(const MatrixT &matrix)
retrieve the size of a square matrix
float Float_t
Definition RtypesCore.h:57
#define R__ASSERT(e)
Definition TError.h:118
void Error(const char *location, const char *msgfmt,...)
Use this function in case an error occurred.
Definition TError.cxx:185
Option_t Option_t TPoint TPoint const char GetTextMagnitude GetFillStyle GetLineColor GetLineWidth GetMarkerStyle GetTextAlign GetTextColor GetTextSize void char Point_t Rectangle_t WindowAttributes_t Float_t Float_t Float_t Int_t Int_t UInt_t UInt_t Rectangle_t Int_t Int_t Window_t TString Int_t GCValues_t GetPrimarySelectionOwner GetDisplay GetScreen GetColormap GetNativeEvent const char const char dpyName wid window const char font_name cursor keysym reg const char only_if_exist regb h Point_t winding char text const char depth char const char Int_t count const char ColorStruct_t color const char Pixmap_t Pixmap_t PictureAttributes_t attr const char char ret_data h unsigned char height h offset
TCpuBuffer GetSubBuffer(size_t offset, size_t start) const
Return sub-buffer of size start starting at element offset.
Definition CpuBuffer.cxx:47
size_t GetSize() const
copy pointer from an external
Definition CpuBuffer.h:102
void CopyTo(TCpuBuffer &) const
Copy data to another buffer.
Definition CpuBuffer.cxx:65
void CopyFrom(const TCpuBuffer &)
Copy data from another buffer.
Definition CpuBuffer.cxx:57
std::shared_ptr< AFloat * > fBuffer
Definition CpuBuffer.h:49
struct TMVA::DNN::TCpuBuffer::TDestructor fDestructor
Class that contains all the data information.
Definition DataSetInfo.h:62
Bool_t IsSignal(const Event *ev) const
UInt_t GetNVariables() const
accessor to the number of variables
Definition Event.cxx:316
Int_t GetNcols() const
TMatrixT.
Definition TMatrixT.h:39
const Int_t n
Definition legend1.C:16
typename std::vector< size_t >::iterator IndexIterator_t
Definition DataLoader.h:42
create variable transformations
void operator()(AFloat **pointer)
Definition CpuBuffer.cxx:30