Age | Commit message (Collapse) | Author | |
---|---|---|---|
2017-01-01 | Revert "TH_GENERIC_USE_HALF=1 by default, half enabled by default"revert-888-halfgeneric | Soumith Chintala | |
2017-01-01 | Merge pull request #888 from torch/halfgeneric | Soumith Chintala | |
TH_GENERIC_USE_HALF=1 by default, half enabled by default | |||
2017-01-01 | TH_GENERIC_USE_HALF=1 by default, half enabled by defaulthalfgeneric | soumith | |
2016-12-31 | cmake C flags fix | Soumith Chintala | |
2016-12-31 | adding TH_SO_VERSION option | Soumith Chintala | |
2016-12-31 | Merge pull request #886 from torch/simdcpuidfix | Soumith Chintala | |
use __get_cpuid when available | |||
2016-12-31 | use __get_cpuid when available | Soumith Chintala | |
2016-12-30 | Merge pull request #875 from atkayu/add_histc2 | Soumith Chintala | |
Add a new function bhistc to calculate histogram of batch of images only once | |||
2016-12-29 | Merge pull request #882 from amrobbins/ppcvectorinstxns | Soumith Chintala | |
Add support for VSX vector instructions on PPC | |||
2016-12-29 | Add support for torch.HalfTensor (#874) | gchanan | |
* Add support for torch.HalfTensor. * Improvements/Simplifications for torch.HalfTensor. Improvements/Simplifications: 1) Defines half type as TH_Half, so as to not conflict with cutorch version. Previously, these were defined as the same "half" type and required proper ordering of includes to ensure type was only defined once, which would have affected all downstream projects. 2) No longer generates math functions that are not actually defined on torch.HalfTensor, e.g. maskedFill, map, etc. 3) Adds tests for all available torch.HalfTensor functions 4) Allows compiling without TH_GENERIC_USE_HALF (so if there's a problem can just unset that in CMakeLists rather than backing out) 5) Some simplifications: removes a new copy optimization and some TH_HALF literal definitions Limitations: Because match functions are not defined, some "non-math" operators on torch.HalfTensor give an error message, e.g. __index__/__newindex__ with a ByteTensor apply a mask, but masks aren't implemented. These limitations aren't always obvious, (e.g. for documentation purposes), but they should always give an error message. * Rename TH_HALF to THHalf. | |||
2016-12-29 | Add support for VSX vector instructions on PPC | Andrew Robbins | |
Added support for the fill, diff, scale, mul and add functions using PPC CPU vector instructions. These are used in place of the versions of these functions written for x86, when compiled on PPC. This fixes a compile failure on PPC | |||
2016-12-28 | fixed the writing error in math doc | atkayu | |
2016-12-28 | rename histc2 to bhistc | atkayu | |
2016-12-22 | add a new function histc2 | atkayu | |
2016-12-20 | Merge pull request #872 from pavanky/modfix | Soumith Chintala | |
correctness fixes for mod and remainder for integer type tensors. | |||
2016-12-20 | correctness fixes for mod and remainder for integer type tensors. | Pavan Yalamanchili | |
2016-12-15 | fix bounds issue in snprintf | soumith | |
2016-12-14 | removing ifdef | soumith | |
2016-12-14 | small fixes to allocator | soumith | |
2016-12-14 | Merge pull request #868 from vfonov/fix_arm | Soumith Chintala | |
Fixed compilation on Raspberry PI without NEON, fixes #807 | |||
2016-12-14 | Fixed compilation on Raspberry PI without NEON | Vladimir S. FONOV | |
2016-11-28 | Merge pull request #851 from BTNC/win | Soumith Chintala | |
2 trivial changes to compile with msvc | |||
2016-11-28 | changes to compile with msvc | Rui Guo | |
2016-11-22 | Merge pull request #845 from killeent/apply-documentation | Soumith Chintala | |
Add some documentation for APPLY and DIM_APPLY macros | |||
2016-11-22 | Add some documentation for APPLY and DIM_APPLY macros | Trevor Killeen | |
2016-11-15 | make the tolerance higher for logNormal test | Soumith Chintala | |
2016-11-15 | Merge pull request #839 from Atcold/fix_ASIMD | Soumith Chintala | |
Fix compilation for ASIMD, fix #766 | |||
2016-11-15 | Fix compilation for ASIMD | Alfredo Canziani | |
On ARMv8, neon is inherit and instead listed as 'asimd' in /proc/cpuinfo Replace assembly with C Original authors: - @dusty-nv FindARM-patch.txt CMakeLists-patch.txt - @rtarquini NEON.c | |||
2016-11-14 | [PATCH] Improve potrf error message. (#189) | Brandon Amos | |
2016-11-11 | Update random.md | Soumith Chintala | |
2016-11-11 | Merge pull request #834 from killeent/lognormal-fix | Soumith Chintala | |
Fix implementation of logNormal | |||
2016-11-11 | Fix implementation of logNormal | Trevor Killeen | |
2016-11-03 | Merge pull request #828 from apaszke/lapack | Soumith Chintala | |
Add more size checks and improve some LAPACK error messages | |||
2016-11-02 | Add more size checks and improve some LAPACK error messages | Adam Paszke | |
2016-11-02 | Merge pull request #827 from howard0su/freebsd | Soumith Chintala | |
Fix compile error on freebsd | |||
2016-11-02 | Fix compile error on freebsd | Howard Su | |
2016-10-27 | Merge pull request #820 from jdmartin86/patch-1 | Soumith Chintala | |
Updated maskedCopy example #1 | |||
2016-10-27 | Updated maskedCopy example #1 | John Martin | |
Original call to maskedCopy() only included y, not the mask. Update added the mask to the first argument. | |||
2016-10-27 | Update .travis.yml | Soumith Chintala | |
2016-10-26 | Merge pull request #819 from colesbury/void | Soumith Chintala | |
Fix no-arg function prototypes | |||
2016-10-26 | Fix no-arg function prototypes | Sam Gross | |
2016-10-26 | Merge pull request #817 from torch/fixmmap | Soumith Chintala | |
fix bug in mmaping | |||
2016-10-26 | fix bug in mmapingfixmmap | soumith | |
2016-10-24 | Merge pull request #812 from howard0su/master | Soumith Chintala | |
Fix the reference so that trepl can find the right document | |||
2016-10-24 | Fix the reference so that trepl can find the right document | Howard Su | |
2016-10-17 | Merge pull request #800 from colesbury/openmp | Soumith Chintala | |
Expose OpenMP num threads through TH lib | |||
2016-10-17 | Expose OpenMP num threads through TH lib | Sam Gross | |
Expose omp_set_num_threads and similar APIs through the TH lib. This means a third-party libaries using TH don't need to be compiled with OpenMP support just to control the number of TH OMP threads. | |||
2016-10-13 | Merge pull request #797 from BTNC/win | Soumith Chintala | |
change timeSort test to work on windows | |||
2016-10-13 | change timeSort test to work on windows | Rui Guo | |
2016-10-13 | Merge pull request #796 from howard0su/master | Soumith Chintala | |
Fix error when run torch.test multi times |