tiny-dnn / tiny-dnn / 1029 / 1
70%
master: 70%

DEFAULT BRANCH: master
Ran
Files 77
Run time 22s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 70.347% (+0.7%) from 69.686%
USE_SSE=OFF USE_AVX=OFF USE_DOUBLE=OFF

push

travis-ci

nyanp 
Tensor (reprise) (#411)

* add Tensor class for CPU mode

* move data pointer to unique_ptr and fix data accessors

* transfer tensor data to device

* rename members, add fromDevice() and accessors methods

* convert to templated class

* refactor constructor and minor changes

* implement make_unique

* fix zero-overhead accessor

* initialize vector with aligned data

* refactor and add more accessor tests

* avoid copy when return shape()

* update fromDevice() routine

* update licence header and fix bound check

* implement fill routine

* implement element-wise add, sub, mul, div

* add move semantics

* add shape checking in element-wise operations

* removing throw exception, adding assert instead

* add std::numeric_limit<U>::min() to avoid zero division

* Add linspace to tensor

* Add sqrt to tensor

* Add exp to tensor

* codestyle

* disable move constructor for VS2013 compilers

* fix typo errors in tests

* fix tensor operations tests to avoid fp related problems

* add more tests to handle tensor shape mismatch

* fix zero division issue

* fix sqrt error for negative numbers

* add destructor,copy ctors, and code cleaning

* remove small value sum in div

* fix tests precision

* workaround solution for unique_ptr in move/copy ctors

* Proposal for tensor

* - renamed private method resize to public reshape
- moved checks to toDevice/fromDevice
- moved operations to non-member functions (layer_*), except fill
- fixed tests

* convert to templated class

* avoid copy when return shape()

* update fromDevice() routine

* implement element-wise add, sub, mul, div

* Proposal for tensor

* Fixing errors revealed by travis

* Fixed VS2012 mistake

* Fixed VS2012 mistake again

3466 of 4927 relevant lines covered (70.35%)

3667511.96 hits per line

Source Files on job 1029.1 (USE_SSE=OFF USE_AVX=OFF USE_DOUBLE=OFF)