-
#Chainer/#CuPy v5.0.0b3をリリースしました。F.einsum, F.lgamma, F.digamma, F.polygammaが追加されました。CuPyではSciPy互換の関数を提供するcupyx.scipy名前空間が追加され、新しいcupy.RawKernelでユーザ定義カーネルをサポートしました。 https://github.com/cupy/cupy/releases/tag/v5.0.0b3 …https://github.com/chainer/chainer/releases/tag/v5.0.0b3 … -
Released
#Chainer/#CuPy v5.0.0b3! New functions: F.einsum, F.lgamma, F.digamma, and F.polygamma. In CuPy, added cupyx.scipy namespace for CUDA-enabled SciPy functions. New user-defined kernel class is added: cupy.RawKernel. https://github.com/cupy/cupy/releases/tag/v5.0.0b3 … https://github.com/chainer/chainer/releases/tag/v5.0.0b3 … -
#Chainer/#CuPy v4.3.0をリリースしました! “Chainer at a Glance” がドキュメントに追加されています: https://docs.chainer.org/en/stable/glance.html …. https://github.com/cupy/cupy/releases/tag/v4.3.0 … https://github.com/chainer/chainer/releases/tag/v4.3.0 … -
Released
#Chainer/#CuPy v4.3.0! Check the new one page tour at https://docs.chainer.org/en/stable/glance.html …. https://github.com/cupy/cupy/releases/tag/v4.3.0 … https://github.com/chainer/chainer/releases/tag/v4.3.0 … -
Released
#Chainer and#CuPy v4.2.0! return_indices option of max pooling functions, broader support of complex arrays, improvement on allocation strategy of pinned memory, and better CuPy ndarray initialization are introduced. https://github.com/chainer/chainer/releases/tag/v4.2.0 …,https://github.com/cupy/cupy/releases/tag/v4.2.0 … -
#Chainer/#CuPy v5.0.0b2 をリリースしました! FP16 モード、確率分布 (今後のリリースでさらに追加予定)、Chainer Functionの自動ブロードキャスト、dilated/grouped (De)ConvolutionND、DLPack と CUDA 9.2 のサポートを行いました。 https://github.com/cupy/cupy/releases/tag/v5.0.0b2 … https://github.com/chainer/chainer/releases/tag/v5.0.0b2 … -
Released
#Chainer/#CuPy 5.0.0b2! FP16 mode, probability distributions (many more coming in next releases), automatic broadcast in Chainer, dilated/grouped (De)ConvolutionND, DLPack & CUDA 9.2 support. https://github.com/cupy/cupy/releases/tag/v5.0.0b2 … https://github.com/chainer/chainer/releases/tag/v5.0.0b2 … -
The PR for DLPack support in CuPy has been merged! Now you can convert CuPy ndarray into PyTorch tensor with zero-copy, and vice versa. You can also convert it into other tensor formats used in libraries which support DLPack too.
#CuPy#DLPack https://github.com/cupy/cupy/pull/1082 …pic.twitter.com/vKsUDEpr7A
-
#Chainer/#CuPy v5.0.0b1 をリリースしました! Iteratorのorder_samplerオプション、LARS、微分可能な疎行列積などが実装され、Windows用のCuPy wheelパッケージが試験的に追加されています。 https://github.com/cupy/cupy/releases/tag/v5.0.0b1 … https://github.com/chainer/chainer/releases/tag/v5.0.0b1 … -
Released
#Chainer/#CuPy 5.0.0b1! order_sampler option of iterators for custom batching constraints, LARS, differentiable sparse matmul, and CuPy wheel package for Windows (experimental). https://github.com/cupy/cupy/releases/tag/v5.0.0b1 … https://github.com/chainer/chainer/releases/tag/v5.0.0b1 … -
Anacondaユーザの皆様、CuPyがcondaコマンドでインストールできるようになりました。"conda install cupy" でビルド済みのCuPyが入ります。(現在はLinux向けのバイナリのみの配布です)
#CuPy -
For Anaconda users, you can now install CuPy as a pre-built binary via the "conda" command: "conda install cupy" (currently only available for Linux)
#CuPy -
Chainer/CuPy v4.0.0をリリースしました!
#Chainer: TensorCore/NCCL2やiDeepのサポートなどによるパフォーマンス向上、Caffeエクスポート機能の試験的な追加など。#CuPy: CUDA9.1をサポート、ビルド済みwheelパッケージを追加、FFT関数の追加など。 https://chainer.org/announcement/2018/04/17/released-v4.html … -
Released Chainer/CuPy v4.0.0!
#Chainer: Major performance improvements including TensorCore support and iDeep backend, NCCL2 support, Caffe export.#CuPy: CUDA 9.1 support, wheel package, FFT support, etc. More in the blog post and release notes. https://chainer.org/announcement/2018/04/17/released-v4.html … -
Released
#CuPy and#Chainer v5 alpha! For CuPy, added “edge” and “reflect” mode for cupy.pad, added cupy.histogram, added “cupyx.scipy.ndimage” for many image transformation like affine/rotate/shift/zoom etc. on GPU! https://github.com/chainer/chainer/releases/tag/v5.0.0a1 … https://github.com/cupy/cupy/releases/tag/v5.0.0a1 … -
4月25日(水)18:00より、Chainer/CuPy v4の新機能紹介とv5に向けたロードマップの紹介、ChainerUI・Chainer Chemistryの紹介を行うイベントを開催します。 場所は大手町ビル(PFNオフィス内多目的室)となります。 奮ってご参加ください。 https://connpass.com/event/59740/
#Chainer#CuPy -
Released the release candidates for CuPy/Chainer v4.0.0!
#CuPy: CUDA9.1 support. You can use it by just running “pip install cupy-cuda91”.#Chainer: FFT, locally-connected 2D layer, etc. The major releases of v4 are planned on Apr. 17. https://github.com/cupy/cupy/releases/tag/v4.0.0rc1 … -
I just met with the creator of
#chainer,@beam2d, and also#cupy, Ryosuke Okuta from@PreferredNetJP. It was so nice to meet these titans of open-source who have created such useful tools. I'm excited to help connect this Pythonic platform to#dask,#kubeflow, and@numba_jit. -
Last week I met with a gracious engineer at
@PreferredNetJP and learned more about#cupy and#chainer. I highly recommend looking at chainer as a more Pythonic machine learning framework: https://chainer.org/ also using https://cupy.chainer.org/ . Think about@numba_jit with it. -
#CuPy#Chainer v4.0.0b4 をリリースしました! CuPyはwheelパッケージにより高速にインストールできるようになりました。ChainerはiDeepを試験的にサポートし、Intel CPU上で性能が向上した他、新しい関数が追加されています。 https://github.com/cupy/cupy/releases/tag/v4.0.0b4 … https://github.com/chainer/chainer/releases/tag/v4.0.0b4 …
লোড হতে বেশ কিছুক্ষণ সময় নিচ্ছে।
টুইটার তার ক্ষমতার বাইরে চলে গেছে বা কোনো সাময়িক সমস্যার সম্মুখীন হয়েছে আবার চেষ্টা করুন বা আরও তথ্যের জন্য টুইটারের স্থিতি দেখুন।