Reduce Mean ONNX Operator implemented #22
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Fix the following compilation error of `ntuple_types` on Windows: ``` ntuple_types.obj : error LNK2019: unresolved external symbol "const ROOT::Experimental::RRecordField::`vftable'" (??_7RRecordField@Experimental@ROOT@@6b@) referenced in function "private: virtual void __thiscall RNTuple_CreateField_Test::TestBody(void)" (?TestBody@RNTuple_CreateField_Test@@EAEXXZ) [C:\Users\bellenot\build\release\tree\ntuple\v7\test\ntuple_types.vcxproj] C:\Users\bellenot\build\release\tree\ntuple\v7\test\Release\ntuple_types.exe : fatal error LNK1120: 1 unresolved externals [C:\Users\bellenot\build\release\tree\ntuple\v7\test\ntuple_types.vcxproj] ```
This reverts commit 71e133d.
This commit reverts root-project/root@69568116cd The CtorsConstants variable is not used. The commit log has a simple tests which now works without this patch. Thanks to Jonas Hahnfeld (@hahnjo) for pointing this out!
Added new main block because this was mentioned in the meta issue list (root-project#406 in cling) as one of the rewrite steps. This also allows for the main code to be run as opposed to running the main code plus the function defintions.
Co-authored-by: Sebastien Binet <binet@cern.ch>
Replace plain Form by TString form; form.Form(...)
Ensure proper allocation, destroy, copy of this data Had problems in copy and assign operators
One cannot copy pointers on pie slices, but have to copy values. Implement TPieSlice::Copy function for that
Probably, is very rare use-case
Simplify cleanup, no need to call `goto L210` only for cleanup. Use nullptr
When custom labels for TGaxis was configured, they were never deleted. Situation is complicated, while fModLabs can point on modified labels from TAxis object. Need to be improved.
Replace `if (flag == kTRUE)` just by `if (flag)` where flag is Bool_t
It is the only workaround to support I/O of TGaxis. If TGaxis was configured from histogram axis and stored, it may store `fModLabs` list from TAxis. After reading back such object from the file one has no information about ownership. Therefore now only lists with configured ownership are deleted - making no crash with old ROOT files.
As per C++ core guidelines.
Before this patch, we were iterating until the end of the map rather than until the end of the equal_range (this was not a problem in practice because program logic guaranteed that we would find what we were looking for and break out of the loop early).
Simplifies memory management
Avoid direct dynamic allocation of plain C arrays
…e-dev-env [skip-ci] (root-project#10960) Added skip-cleanup flags for create-dev-env. This would instruct the cpt to build cling, clang, and llvm, without deleting the build area, as the cpt normally would delete the build area due to the use of memory running out when linking, because of too many build subprocesses.
Header was badly formatted which resulted in only part of it being displayed. Also, fixed some typos.
This is to suppress the warnings we see now in the nightlies on the macbeta nodes: https://lcgapp-services.cern.ch/root-jenkins/view/ROOT%20Nightly/job/root-nightly-master/LABEL=macbeta,SPEC=cxx17,V=master/lastBuild/parsed_console/
The RooWorkspace sometimes stores some RooArgSet prefixed with `CACHE_` in itself to cache for example parameters or constraint sets. These cache sets are invalidated when elements are removed from them by `RooWorkspace::RecursiveRemove()`. In this case, they should be removed from the workspace such that they can be correctly recomputed later. This change fixes problems like this one reported on the forum: https://root-forum.cern.ch/t/how-to-properly-redefine-pdf-in-rooworkspace/50757 In that usecase, the following sequence of events happened: 1. Create pdf in workspace 2. Fit this pdf (triggering the caching of the set of parameters) 3. Recursively remove everything from the workspace, but not the cache sets as they are hidden from the user. The cached parameter sets are now empty, as all parameters got removed from the workspace 4. The same pdf from step 1 is recreated 5. Fitting of this new pdf will now fail, because RooFit thinks it has zero free parameters, as the call to `getParameters()` now returns the cleared cache set!
Some modernization of the RooWorkspace code: * use less `0` to represent null pointers and booleans * use more `!empty()` instead of `size()>0` * replace `MakeIterator` with range-based loops * remove some dead code that was commented out for over a decade
So far, one had to remove string attributes of a RooAbsArg with the `RooAbsArg::removeStringAttribute(const char* key, const char* value)` method, passing `nullptr` as a value. Passing a `nullptr` to a function that takes a `const char*` is not supported in PyROOT, as explained in GitHub issue root-project#10954. Therefore, a new function is added to RooAbsArg to remove a string attribute without having to use `nullptr`. This new function called `removeStringAttribute()` is now used in all the ROOT code to remove string attributes from RooAbsArgs, and also mentioned in the warnings that tell the user that they might want to remove the `fitrange` attribute. The verb `remove` was chosen instead of `delete`, `clear`, or `erase`, because there was already a `RooAbsArg::removeServer()` method.
This is necessary for getting the code to be flake8 compliant.
Some of the quotes are causing issues in some `autoconf` macros for two packages in the LCG stack
Generalizes `RPageSourceDaos::LoadClusters()` to batch up the fetch requests for all pages in the cluster bunch before calling `RDaosContainer::ReadV`. Async fetch requests for pages from different clusters may now share queue and flight time; thus, remote storage can parallelize them and improve read throughput.
Last year, with commit 3657e7c, the parameter index calculation was changed to be on the fly instead of using a look-up map, which is much faster. However, the implemented formula was not correct for two or three dimensions, which is fixed by this commit. To make sure that the index computation is correct this time, the new code was tested in this code snippet with various inputs: ```C++ void runTest(int nx = 42, int ny = 42, int nz = 42) { const int nxy = nx * ny; const int nyz = ny * nz; for (int i = 0; i < nx; ++i) { for (int j = 0; j < ny; ++j) { for (int k = 0; k < nz; ++k) { const int index = k + j * nz + i * ny * nz; const int gammaIndex = i + j * nx + k * nx * ny; const int i2 = index / nyz; const int tmp = index % nyz; const int j2 = tmp / nz; const int k2 = tmp % nz; const int gammaIndex2 = i2 + j2 * nx + k2 * nxy; if (gammaIndex2 != gammaIndex) { std::cout << "The unraveled indices were not correct!" << std::endl; return; } } } } } ``` Needs to be backported to the 6.26 branch to get into the 6.26.06 patch release. This commit the following problem reported on the forum: https://root-forum.cern.ch/t/cpycppyy-segfault-on-mac-m1/50822
Enables Barlow-Beeston in multidimensional fits. This patch was extensively tested and used by P. Hamilton et al. in LHCb. Loop over all entries, not just the x-axis of a histogram.
This is to suppress the warnings we see now in the nightlies on the macbeta nodes: https://lcgapp-services.cern.ch/root-jenkins/view/ROOT%20Nightly/job/root-nightly-master/LABEL=macbeta,SPEC=cxx17,V=master/lastBuild/parsed_console/ After this PR, there will only be a final PR necessacy to fix these warnings in TMVA.
When the normalization range for coefficient determination of a RooAddPdf is changed, the AddPdf's projection cache needs to be reset, just like it is already done in `RooAddPdf::fixCoefNormalization`. Otherwise, there will be problems in the pdf evaluation and integration because the projection cache is invalid. A unit test based on the GitHub issue that reported this problem is also implemented. Closes root-project#10988.
It should be avoided to have RooAbsArgs with the same object name as part of the same computation graph. This has happened in RooAddPdf when using internally defined recursive coefficients, which is changed with this commit.
TMVA DataSetInfo incorrectly handles array of variables
Fixed formatting issues
Fixed formatting issues
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.