Changelog History
-
v1.0.0.a3 Changes
November 29, 2016๐ Now we are announcing v1.0.0a3. Thanks for all great contributors! This release includes the following changes from v1.0.0a2:
๐ Bug fix
- ๐ Convolutional layer with
padding::same
mode doesn't work #332 fixed by @nyanp - ๐ Segmentation fault at MinGW build #203 #281 fixed by @nyanp
- ๐ NNPACK backend doesn't work #398 fixed by @azsane
๐ Improvements
- โ Remove compiler warnings & improve CMakeLists #387 by @beru
- ๐ Improve memory consumption #410 by @beru
- ๐ Improve unit tests #408 by @Randl
- Subtle speed optimization #419 by @beru
- ๐จ Refactor serialization type & size type #407, #422 by @Randl and @edgarriba
- ๐ Improve compilation time by splitting serialiation/deserialization #421 by @beru
๐ Docs&Comments
- โ Add comments to layer class #424 by @edgarriba
- ๐ Fix typo in comments #404 by @MikalaiDrabovich
Toward v1.0.0
๐ The first version of
tensor
is merged into tiny-dnn (#411 #417 #418 by @pansk @Randl and @edgarriba). It isn't integrated with tiny-dnn layers yet, but it's the starting point of the GPU tiny-dnn. - ๐ Convolutional layer with
-
v1.0.0.a2 Changes
November 13, 2016๐ Bug Fixes
- ๐ Fix SEGV errors on AVX Optimized code (#353) by @nyanp
- ๐ Fix compiler error on msvc2013 (#320) by @nyanp
- ๐ Fix AVX backend slowdown on convolutional layer (#322) by @nyanp
- ๐ Fix throwing error when we load weights manually (#330) by @nyanp
- ๐ Fix returning infinity in tan_h (#347) by @nyanp
- ๐ Fix portability issues on serialization (#377) by @nyanp
๐ Features
- ๐ Provides compile option to disable serialization support to speedup compilation time (#316) by @nyanp
- โ Adds
set_trainable
method to freeze layers (#346) by @nyanp - โ Adds power layer to caffe converter by @goranrauker
- ๐ double precision support (#332) by @nyanp
- Provides pad_type and non-square input to pooling layers (#374) by @nyanp
- โ Adds public predict method for vector of tensors (#396) by @reunanen
- โ Adds Auto engine selection (#339) by @edgarriba
- โ Adds basic image utilities, and remove OpenCV dependencies (#337) by @nyanp
Others
- ๐ Sync with latest NNPACK by @azsane
- ๐ Improves compiler warnings around type-cast by @pansk @reunanen @edgarriba
- ๐ Improves CMakelist by @syoyo @edgarriba @beru
- โ Replaces picotest with gtest by @Randl
- โ Adds "layer catalogue" into official documentation by @nyanp
- โ Adds tests for GPU environment by @Randl
- โ Adds cpplint.py by @edgarriba @Randl
- โ Adds a document for building iOS app by @wangyida
- โ Adds coverall checking by @edgarriba
- โ Adds CI builds for Win32 by @nyanp
- โก๏ธ Updates & improves readme by @edgarriba@zhangqianhui
-
v1.0.0.a Changes
September 14, 2016๐ ๐ This release contains a major refactoring & many bugfixes. Thanks a lot for all great contributors! ๐
๐ This release is alpha version. We need more helps and feedbacks toward v1.0.0. Please submit your bug-report at Github issue. Many thanks :)
โก๏ธ Major updates
- Merge successful results from GSoC 2016
These features are still experimental, so PRs and bug reports are very welcome!
- Model serialization by @nyanp
Minor bug fix
- Memory errors in cifar-10 example #295 #300 by @edgarriba
- Fix max-pooling layer #271 by @nyanp
- Fix concat-layer #301 by @Jiaolong
- Suppress compiler warnings #297 by @syoyo
Other
- A nice project logo by @KonfrareAlbert
- Launch official documents at http://tiny-dnn.readthedocs.io/
some APIs are changed from v0.1.1 .
- changed its namespace from
tiny_cnn
totiny_dnn
- changed API header from
tiny_cnn.h
totiny_dnn.h
-
v0.1.1 Changes
July 26, 2016 -
v0.1.0 Changes
June 06, 2016๐ This release contains a major refactoring around fundamental architecture of tiny-cnn and fixes many problems. We had the help of 20 comitters for this release. Thanks!
- Now we can handle non-sequential model as
network<graph>
#108
#153 - โ Catch up the latest format of caffe's proto #162
- ๐ Improve the default behaviour of re-init weight #136
- โ Add more tests and documents #73
- โ Remove dependency of OpenCV in MNIST example
๐ Some API have changed from the previous release. see change list
- Now we can handle non-sequential model as
-
v0.0.1
May 22, 2016