Nnabla-ext-cuda

Latest version: v1.30.0

Safety actively analyzes 628918 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 3 of 6

1.21.0

release-note-break-compat
+ [Make inplace option in most of Function operations obsolete](https://github.com/sony/nnabla-ext-cuda/pull/336)

release-note-bugfix
+ [Re-enable the cudnn convolution metrics caching.](https://github.com/sony/nnabla-ext-cuda/pull/327)
+ [Fix multithread cudnn handle sharing problem ](https://github.com/sony/nnabla-ext-cuda/pull/324)

release-note-build
+ [Install nnabla_converter to py38 docker images](https://github.com/sony/nnabla-ext-cuda/pull/337)
+ [Fix APT_OPTS for runtime dockerfiles](https://github.com/sony/nnabla-ext-cuda/pull/332)
+ [CI: Support Nvidia Compatible Driver for CUDA11 docker images](https://github.com/sony/nnabla-ext-cuda/pull/328)

release-note-doc
+ [update copyright](https://github.com/sony/nnabla-ext-cuda/pull/325)

release-note-op-layer
+ [Modify cumprod backward cuda version to deal with zero input](https://github.com/sony/nnabla-ext-cuda/pull/338)
+ [Support p=0 for dropout function](https://github.com/sony/nnabla-ext-cuda/pull/335)
+ [ignore label less 0 to support segmentation boundary](https://github.com/sony/nnabla-ext-cuda/pull/334)
+ [Optimize SyncBatchNorm](https://github.com/sony/nnabla-ext-cuda/pull/333)
+ [ Boolean Indexing Functions](https://github.com/sony/nnabla-ext-cuda/pull/331)
+ [Remove ReLUCudaCudnn](https://github.com/sony/nnabla-ext-cuda/pull/329)
+ [Add CUDA implementaion of roi_align function.](https://github.com/sony/nnabla-ext-cuda/pull/326)
+ [ Higher-order gradients for normalization functions](https://github.com/sony/nnabla-ext-cuda/pull/323)
release-note-utility
+ [Implement module nvml to replace pynvml](https://github.com/sony/nnabla-ext-cuda/pull/322)

Install the latest nnabla by:

pip install nnabla
pip install nnabla-ext-cuda102 For CUDA version 10.2 users


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)

1.20.1

release-note-bugfix
+ [install flatbuffer compiler for new tflite converter](https://github.com/sony/nnabla-ext-cuda/pull/320)

Install the latest nnabla by:

pip install nnabla
pip install nnabla-ext-cuda102 For CUDA version 10.2 users


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)

1.20.0

release-note-bugfix
+ [Bug fix of fused binary operator](https://github.com/sony/nnabla-ext-cuda/pull/316)
+ [ Bugfix: forward(clear_no_need_grad=True) and nn.grad of Dropout](https://github.com/sony/nnabla-ext-cuda/pull/313)
+ [Fix buffer overrun with cudnnBatchNormalizationForwardTraining](https://github.com/sony/nnabla-ext-cuda/pull/312)
+ [Fix random number range in RandomCropCuda](https://github.com/sony/nnabla-ext-cuda/pull/310)
+ [Fix issue with softplus](https://github.com/sony/nnabla-ext-cuda/pull/309)

release-note-build
+ [add the support of copyright check](https://github.com/sony/nnabla-ext-cuda/pull/315)
+ [version-up of cuda and python](https://github.com/sony/nnabla-ext-cuda/pull/314)

release-note-core
+ [Support forward recomputation in backward propagation](https://github.com/sony/nnabla-ext-cuda/pull/311)

release-note-doc
+ [Fix jupyter command](https://github.com/sony/nnabla-ext-cuda/pull/306)

release-note-op-layer
+ [Modify SoftPlus to accept beta argument](https://github.com/sony/nnabla-ext-cuda/pull/308)
+ [Searchsorted and Meshgrid implementation](https://github.com/sony/nnabla-ext-cuda/pull/304)

release-note-utility
+ [fix: pynvml.nvml has no attribute nvml_lib](https://github.com/sony/nnabla-ext-cuda/pull/317)
+ [Use urllib instead of requests](https://github.com/sony/nnabla-ext-cuda/pull/307)

Install the latest nnabla by:

pip install nnabla
pip install nnabla-ext-cuda102 For CUDA version 10.2 users


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)

1.19.0

release-note-bugfix
+ [Change workspace limit dtype from int to int64_t](https://github.com/sony/nnabla-ext-cuda/pull/298)
+ [fix conv segment fault error](https://github.com/sony/nnabla-ext-cuda/pull/296)
+ [Fix numerical instability for log_sigmoid and softplus](https://github.com/sony/nnabla-ext-cuda/pull/295)

release-note-build
+ [remove conda](https://github.com/sony/nnabla-ext-cuda/pull/297)

release-note-core
+ [clears input buffer not needed for after propagation](https://github.com/sony/nnabla-ext-cuda/pull/299)

release-note-op-layer
+ [fused binary operator](https://github.com/sony/nnabla-ext-cuda/pull/300)
+ [F.cumsum and F.cumprod](https://github.com/sony/nnabla-ext-cuda/pull/294)

Install the latest nnabla by:

pip install nnabla
pip install nnabla-ext-cuda102 For CUDA version 10.2 users


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)

1.18.0

release-note-bugfix

+ [fix: solve the error that pynvml cannot solve the path to nvml.dll on windows](https://github.com/sony/nnabla-ext-cuda/pull/293)
+ [Load bias grad only if convolution has bias.](https://github.com/sony/nnabla-ext-cuda/pull/292)
+ [fix key error of checking GPU](https://github.com/sony/nnabla-ext-cuda/pull/291)

release-note-core

+ [Higher-order Gradients](https://github.com/sony/nnabla-ext-cuda/pull/290)

release-note-op-layer

+ [Memory optimization for SpectralNorm](https://github.com/sony/nnabla-ext-cuda/pull/289)

Install the latest nnabla by:

pip install nnabla
pip install nnabla-ext-cuda102 For CUDA version 10.2 users


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)

1.17.0

release-note-bugfix
+ [Fix a bug in LSTM](https://github.com/sony/nnabla-ext-cuda/pull/288)
+ [enable multi-thread training](https://github.com/sony/nnabla-ext-cuda/pull/277)

release-note-build
+ [Limit numpy version to 1.19.5](https://github.com/sony/nnabla-ext-cuda/pull/287)
+ [Add cudnn libraries to whl](https://github.com/sony/nnabla-ext-cuda/pull/283)

release-note-core
+ [Add seed function](https://github.com/sony/nnabla-ext-cuda/pull/279)

release-note-op-layer
+ [Uses heuristics to decide conv algorithm as default](https://github.com/sony/nnabla-ext-cuda/pull/286)
+ [Split convolution workspace size](https://github.com/sony/nnabla-ext-cuda/pull/285)
+ [Deformable Convolution](https://github.com/sony/nnabla-ext-cuda/pull/284)
+ [Memory optimization for STFT/iSTFT](https://github.com/sony/nnabla-ext-cuda/pull/282)
+ [Optimize normalization functions implementation](https://github.com/sony/nnabla-ext-cuda/pull/281)
+ [Pad memory optimization and add repeat mode](https://github.com/sony/nnabla-ext-cuda/pull/280)

release-note-utility
+ [check gpu compatibility](https://github.com/sony/nnabla-ext-cuda/pull/278)

Install the latest nnabla by:

pip install nnabla
pip install nnabla-ext-cuda102 For CUDA version 10.2 users


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)

Page 3 of 6

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.