Nnabla-ext-cuda

release-note-bugfix

+ [Fix a race condition when using thrust::async::sort](https://github.com/sony/nnabla-ext-cuda/pull/188)
release-note-build

+ [Exit with 0 if no tests in pytest](https://github.com/sony/nnabla-ext-cuda/pull/186)

release-note-core

+ [Improve the speed of transpose function](https://github.com/sony/nnabla-ext-cuda/pull/190)

release-note-examples

+ [Added auto-forward for pycpp](https://github.com/sony/nnabla-ext-cuda/pull/187)

release-note-op-layer

+ [Use thrust async sort and custom make_seqeuence for faster operation.](https://github.com/sony/nnabla-ext-cuda/pull/185)

release-note-utility

+ [Add DaliIterator](https://github.com/sony/nnabla-ext-cuda/pull/189)


Install the latest nnabla by:


pip install nnabla
pip install nnabla-ext-cuda101 For CUDA version 10.1 users


Note that CUDA 10.1 and cuDNN 7.6 are set as default if versions are not specified. You can also install the cuda extension with specific versions from one of the following. [See also FAQ](https://nnabla.readthedocs.io/en/latest/python/install_on_linux.html#faq)

release-note-op-layer

+ [Use thrust async sort and custom make_seqeuence for faster operation.](https://github.com/sony/nnabla-ext-cuda/pull/185)
* **Known issue**: it causes segfault due to an improper concurrency in CUDA 10.1, which should be fixed in the next release
+ [Update FlipCuda and RandomFlipCuda implementation](https://github.com/sony/nnabla-ext-cuda/pull/184)
+ [Min-max quantization](https://github.com/sony/nnabla-ext-cuda/pull/183)

Install the latest nnabla by:


pip install nnabla
pip install nnabla-ext-cuda101 For CUDA version 10.1 users


Users with python <= 3.4 may experience errors with `pip install nnabla` and `pip install nnabla-ext-cuda`.
■ Workaround
Please install matplotlib == 2.2.3 and re-install nnabla, nnabla-ext-cuda.

pip install matplotlib==2.2.3
pip install nnabla
pip install nnabla-ext-cuda101


Note that CUDA 10.1 and cuDNN 7.6 are set as default if versions are not specified. You can also install the cuda extension with specific versions from one of the following. [See also FAQ](https://nnabla.readthedocs.io/en/latest/python/install_on_linux.html#faq)

release-note-bugfix

+ [fix convolution segv caused by cudaEvent_t instance variables](https://github.com/sony/nnabla-ext-cuda/pull/177)
+ [Generate random seed when seed==-1 in cudnn context](https://github.com/sony/nnabla-ext-cuda/pull/176)

release-note-build

+ [Includes cuda lib in wheel](https://github.com/sony/nnabla-ext-cuda/pull/182)

release-note-op-layer

+ [Add CUDA implementation of OneHot function](https://github.com/sony/nnabla-ext-cuda/pull/180)
+ [RandomFlip implementation in CUDA](https://github.com/sony/nnabla-ext-cuda/pull/179)

release-note-utility

+ [Skip allreduce if array is not updated](https://github.com/sony/nnabla-ext-cuda/pull/181)


Install the latest nnabla by:


pip install nnabla
pip install nnabla-ext-cuda101 For CUDA version 10.1 users


Users with python <= 3.4 may experience errors with `pip install nnabla` and `pip install nnabla-ext-cuda101`.
■ Workaround
Please install matplotlib == 2.2.3 and re-install nnabla, nnabla-ext-cuda.

pip install matplotlib==2.2.3
pip install nnabla
pip install nnabla-ext-cuda101


Note that CUDA 10.1 and cuDNN 7.6 are set as default if versions are not specified. You can also install the cuda extension with specific versions from one of the following. [See also FAQ](https://nnabla.readthedocs.io/en/latest/python/install_on_linux.html#faq)