Package: caffe-nv Architecture: ppc64el Version: 0.14.5-3ibm1 Priority: optional Section: science Maintainer: Anup Halarnkar Installed-Size: 77886 Depends: build-essential, cmake, curl, g++, gcc, git-core, protobuf-compiler, python (>= 2.7), python-skimage, python-protobuf, cuda-cublas-8-0, cuda-cudart-8-0, cuda-curand-8-0, libboost-python1.58.0, libboost-system1.58.0, libboost-thread1.58.0, libc6 (>= 2.17), libcudnn5-dev, libgcc1 (>= 1:3.0), libgflags2v5, libgoogle-glog0v5, libhdf5-10, libleveldb1v5, liblmdb0 (>= 0.9.7), libopenblas, libopencv-core2.4v5, libopencv-highgui2.4v5, libopencv-imgproc2.4v5, libprotobuf9v5, libpython2.7 (>= 2.7), libstdc++6 (>= 5.2) Conflicts: caffe Filename: dists/xenial/main/binary-ppc64el/caffe-nv_0.14.5-3ibm1_ppc64el.deb Size: 12214284 MD5sum: 70d9ab727c15277103db043076933434 SHA1: 636076d1e9a5859fb09543244cd19cb3e9c1bd99 SHA256: 63f537c6f1175693df96433144bb2ade6b83e43f2bfaec30d613145909b28393 SHA512: 75185c662e7bfe171f8d82d495ad9f26897214ec48ffd1ebbe749e53dc8b0cd14e276062564490f47f8313d15022b0bcdf2ff38c46d6f0c4a064a154a31714a7 Homepage: http://caffe.berkeleyvision.org/ Description: Fast open framework for deep learning (NVIDIA's fork) Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and by community contributors. . This package includes NVIDIA's fork of upstream Caffe. Package: python-flask-socketio Architecture: all Version: 2.7.2-3ibm1 Priority: optional Section: python Maintainer: Frederic Bonnard Installed-Size: 67 Depends: python-engineio, python-flask (>= 0.9), python-socketio-server, python:any (<< 2.8), python:any (>= 2.7.5-5~) Filename: dists/xenial/main/binary-ppc64el/python-flask-socketio_2.7.2-3ibm1_all.deb Size: 14512 MD5sum: f8e8751cf352afae48e86ce230b0a44f SHA1: 32662d3056edea1e0abe62f40d49044b39e6a455 SHA256: 940d4fe92db448f8fa4d7a8e5c7dfa2d5eba84724fd951ee6a5cfc5880111189 SHA512: 5426a7302b08bf191addab4fe083471a4f5c166aa4b801f61c643d8fc748027401c098be9a37c256797f351f92ee231eda48d170de22c3ba86f85e553f42dca4 Homepage: https://github.com/miguelgrinberg/Flask-SocketIO Description: Python Socket.IO integration for Flask applications (Python 2) Flask-SocketIO gives Flask applications access to low latency bi-directional communications between the clients and the server. The client-side application can use any of the SocketIO official clients libraries in Javascript, C++, Java and Swift, or any compatible client to establish a permanent connection to the server. . This package contains the module for Python 2. Package: torch Architecture: ppc64el Version: 7-4ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 922948 Depends: gnuplot-x11, gnuplot, graphicsmagick, imagemagick, libgraphicsmagick1-dev, ipython, libfftw3-single3, python-zmq, sox, cuda-cublas-8-0, cuda-cudart-8-0, cuda-cusparse-8-0, libc6 (>= 2.22), libgcc1 (>= 1:3.3), libgomp1 (>= 4.9), libjpeg8 (>= 8c), liblmdb0 (>= 0.9.7), libopenblas, libpng12-0 (>= 1.2.13-4), libqt4-svg (>= 4:4.5.3), libqtcore4 (>= 4:4.8.4), libqtgui4 (>= 4:4.6.1), libreadline6 (>= 6.0), libssl1.0.0 (>= 1.0.0), libstdc++6 (>= 5.2), libx11-6, libzmq5 (>= 4.1.2) Recommends: ipython-notebook Filename: dists/xenial/main/binary-ppc64el/torch_7-4ibm1_ppc64el.deb Size: 47024668 MD5sum: f3cfe205aa214af6dc84cb267747d367 SHA1: 0634fbafb4549605c675b14c3055cc3e79049f73 SHA256: 36b14f1f7524bfde960b5b0c1e6e025217ee46fc1424127c9e0ecdf58545d3ec SHA512: 9e036202cd2565ea2543f61145d0306d277d69826266df979418dabb85e07bad9a8fd150899b0380bdae80b64c89b6478c3f4c4e78110bc011badaa7e2adafdb Homepage: http://torch.ch Description: A scientific computing framework for LuaJIT Torch is a scientific computing framework with wide support for machine learning algorithms that puts GPUs first. It is easy to use and efficient, thanks to an easy and fast scripting language, LuaJIT, and an underlying C/CUDA implementation. Package: python3-flask-socketio Architecture: all Version: 2.7.2-3ibm1 Priority: optional Section: python Source: python-flask-socketio Maintainer: Frederic Bonnard Installed-Size: 67 Depends: python3-engineio, python3-flask (>= 0.9), python3-socketio-server, python3:any (>= 3.3.2-2~) Filename: dists/xenial/main/binary-ppc64el/python3-flask-socketio_2.7.2-3ibm1_all.deb Size: 14590 MD5sum: 0b2d97e27744b2c3776af5008f608e2f SHA1: 1a4c2989092da3f69afff8cf587512b79997357c SHA256: 4c1609d29633d4fc95d0b9e27ae762b1d3de052474ff8f08d3383057654743ae SHA512: 1ef115f919d89804d246f1dcdf125f9781b77f5b23f639e8332157997a42437f538afdbced14a69dd2e65acb775763c52b146ae983ff6f28aa6c7850d4517f3e Homepage: https://github.com/miguelgrinberg/Flask-SocketIO Description: Python Socket.IO integration for Flask applications (Python 3) Flask-SocketIO gives Flask applications access to low latency bi-directional communications between the clients and the server. The client-side application can use any of the SocketIO official clients libraries in Javascript, C++, Java and Swift, or any compatible client to establish a permanent connection to the server. . This package contains the module for Python 3. Package: libnccl1 Architecture: ppc64el Version: 1.3.2-1.cuda8.0 Priority: optional Section: libs Source: nccl Maintainer: cudatools Installed-Size: 23017 Depends: cuda-cudart-8-0, libc6 (>= 2.17), libgcc1 (>= 1:3.0), libstdc++6 (>= 4.1.1) Filename: dists/xenial/main/binary-ppc64el/libnccl1_1.3.2-1.cuda8.0_ppc64el.deb Size: 1318884 MD5sum: 295e5261c029fa116a0ab5bb49fc8ffa SHA1: 29b940dd8a6894680533dfdd70b4ecd09e8a507c SHA256: 08c40a185558356a310b62c482731110678549d2986278a4713846b289151350 SHA512: fac03ad08ba522e7e447f38634d75cfab351c98c5c5c154fbe85670f921d3c0bd6178aeb0497c36fddb24fe54f0afe9335cf485b32f06572950da42bbed1ce3d Description: NVIDIA Collectives Communication Library (NCCL) Runtime NCCL (pronounced "Nickel") is a stand-alone library of standard collective communication routines for GPUs, such as all-gather, reduce, broadcast, etc., that have been optimized to achieve high bandwidth over PCIe. NCCL supports up to eight GPUs and can be used in either single- or multi-process (e.g., MPI) applications. Package: python3-scikit-fmm Architecture: ppc64el Version: 0.0.9-3ibm1 Priority: optional Section: python Source: scikit-fmm Maintainer: Frederic Bonnard Installed-Size: 202 Depends: python3-numpy (>= 1:1.10.0~b1), python3-numpy-abi9, python3 (<< 3.6), python3 (>= 3.5~), libc6 (>= 2.17), libgcc1 (>= 1:3.0), libstdc++6 (>= 5.2) Suggests: python-scikit-fmm-doc Filename: dists/xenial/main/binary-ppc64el/python3-scikit-fmm_0.0.9-3ibm1_ppc64el.deb Size: 42246 MD5sum: a3bbcaf4e628d0240b7f5633fbf3deb6 SHA1: d3b58b1ffcfbba24cd9be4648ce843844bacde5b SHA256: c1411f9e7498f312651d53b41f3e9ddf31c5b13464a5959a1c96307c549db894 SHA512: 5cda42d3d53f4503194ee637d0e1d059e847969afd81a9b4e70d3718150d62a7f3f341570a1211d134c1337b8f91c562ff957f3ea899d96e2f45778a42978d83 Homepage: https://github.com/scikit-fmm/scikit-fmm Description: Fast marching method for Python (Python 3) scikit-fmm is a Python extension module which implements the fast marching method. The fast marching method is used to model the evolution of boundaries and interfaces in a variety of application areas. More specifically, the fast marching method is a numerical technique for finding approximate solutions to boundary value problems of the Eikonal equation. . This package contains the module for Python 3. Package: power-mldl Architecture: ppc64el Version: 3.4.1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 4 Depends: bazel (>= 0.4.3-3ibm1), caffe-bvlc (>= 1.0.0rc5-3ibm1), caffe-ibm (>= 1.0.0rc3-3ibm4), caffe-nv (= 0.14.5-3ibm1) | caffe-nv (>= 0.15.14-3ibm1), chainer (>= 1.20.0.1-3ibm1), digits (>= 5.0-3ibm4), libopenblas (>= 0.2.19-3ibm2), libnccl1 (>= 1.3.2-1.cuda8.0), libnccl-dev (>= 1.3.2-1.cuda8.0), tensorflow (= 0.12.0-3ibm1) | tensorflow (>= 1.0.1-3ibm1), theano (>= 0.9.0-3ibm1), torch (>= 7-3ibm3) Filename: dists/xenial/main/binary-ppc64el/power-mldl_3.4.1_ppc64el.deb Size: 2346 MD5sum: b4c98fa50643be7160609d78ea2ebfab SHA1: 67d2030ac4775dc8be5de5783791bd70791458de SHA256: 008d01da8f6768a7612e4452389bbc209e83b33727dc678e22482c6f31f681dd SHA512: 55125f9f8180928cf1f35a370b9e5a7c04f06ebe0a2bc23750a3c3709f544108a61a517e2f17a0b4354d48752bdf8bd2705d9b1aabfe24dc27a714802cd8224b Description: Meta-package for Deep Learning frameworks for POWER This package does not include any Deep Learning binaries. It is a convenience to allow easy install of the Deep Learning frameworks currently packaged by IBM. Package: theano Architecture: ppc64el Version: 0.9.0-3ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 34809 Depends: libopenblas, python (>= 2.7), python-nose, python-numpy (>= 1.9.1), python-scipy (>= 0.14), python-mako (>= 0.7), python-six, g++, libc6 (>= 2.17) Recommends: python-flake8, python-pydot Filename: dists/xenial/main/binary-ppc64el/theano_0.9.0-3ibm1_ppc64el.deb Size: 10333510 MD5sum: 430d8bd94624e98b258aad71144e5f75 SHA1: 81f5a43c4ff1284cf57b492238fe74dd98109dc2 SHA256: f28ecacb219526a380a483fdd91fb8c337f53503a7f554625f5a4b5397f56873 SHA512: 983e6132a99f65d7be57f1ab414e280250f6f8f9332527d0c9e7bceb6cd073f1603cef716eb6506c56d4abc173012743cf6c504fb56b0cb31c933a7db829d33a Homepage: http://www.deeplearning.net/software/theano/ Description: A Python library for Deep Learning Theano is a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently. This package includes support for NVIDIA GPUs, and requires NVIDIA CUDA Toolkit and CuDNN. Package: digits Architecture: ppc64el Version: 5.0-3ibm4 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 17565 Depends: build-essential, caffe-nv | caffe-bvlc | caffe-ibm, python-flask (>= 0.10.1), python-flask-socketio (>= 2.6.0), python-flaskext.wtf (>= 0.11), python-wtforms (>= 2.1), python-six (>= 1.5.2), python-requests (>= 2.2.1), python-setuptools (>= 3.3), python-eventlet (>= 0.13.0), python-pil (>= 2.3.0), python-numpy (>= 1.8.2), python-scipy (>= 0.13.3), python-protobuf (>= 2.5.0), python-lmdb (>= 0.87), python-gevent (>= 1.0), python-gevent-websocket (>= 0.9.3), python-h5py (>= 2.2.1), python-pydot, python-skimage, python-nose (>= 1.3.1), python-bs4 (>= 4.2.1), python-mock (>= 1.0.1), python-coverage (>= 3.7.1), python-selenium (>= 2.25.0), python-matplotlib (>= 1.3.1), python-psutil (>= 3.4.2), python-scikit-fmm (>= 0.0.9), python (>= 2.7), libhdf5-serial-dev Recommends: torch Conflicts: python-socketio Filename: dists/xenial/main/binary-ppc64el/digits_5.0-3ibm4_ppc64el.deb Size: 10060476 MD5sum: f8550acf4251e9ed0d06362b5e2ef242 SHA1: ca1abc7deee2ea4542256590c4cc0b8b17c2197d SHA256: 341bdbd0e807842434c266eeb2975213e1a35e098afe69babc40fd0d32c3e218 SHA512: c3707379638c8be6422fd5fcf90aacc0bdb27ca398e0a40b8b7ea25dd686e8c883cc5cb9b9588eda2f932460618a1cf590e1445cf76d7256b869602f143e8a45 Homepage: https://developer.nvidia.com/digits Description: Deep Learning GPU Training System DIGITS (the Deep Learning GPU Training System) is a webapp for training deep learning models. Package: python-socketio-server Architecture: all Version: 1.6.0-3ibm1 Priority: optional Section: python Maintainer: Frederic Bonnard Installed-Size: 79 Depends: python-engineio, python-six (>= 1.9.0), python:any (<< 2.8), python:any (>= 2.7.5-5~) Filename: dists/xenial/main/binary-ppc64el/python-socketio-server_1.6.0-3ibm1_all.deb Size: 16426 MD5sum: ab297fc8e42980098517df9d93290141 SHA1: 0ef64d05b63e75258efe8350e5522f2a8fb51be1 SHA256: 5abfea666fde4779f6ac94d92cd3d989edb90e1ca67a95ed294ce64b0efc4eb9 SHA512: 093a925c5dc85d5e998c3e25bd39e6de08e6fd46181e21d076db9a369f1f9a652045011300ff3417bec7c454da2ba9199125bfe05d9fcb1d7f0d53837d7fa3d5 Homepage: https://github.com/miguelgrinberg/python-socketio Description: Python implementation of the Socket.IO realtime server (Python 2) python-engineio project implements a Socket.IO server that can run standalone or integrated with a Python WSGI application. Socket.IO is a transport protocol that enables real-time bidirectional event-based communication between clients (typically web browsers) and a server. The original implementations of the client and server components are written in JavaScript. . This package contains the module for Python 2. Package: caffe-ibm Architecture: ppc64el Version: 1.0.0rc3-3ibm4 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 92048 Depends: build-essential, cmake, curl, g++, gcc, git-core, protobuf-compiler, python (>= 2.7), python-skimage, python-protobuf, cuda-cublas-8-0, cuda-cudart-8-0, cuda-curand-8-0, libboost-filesystem1.58.0, libboost-python1.58.0, libboost-system1.58.0, libboost-thread1.58.0, libc6 (>= 2.17), libcudnn5-dev, libgcc1 (>= 1:3.0), libgflags2v5, libgoogle-glog0v5, libhdf5-10, libleveldb1v5, liblmdb0 (>= 0.9.7), libopenblas, libopencv-core2.4v5, libopencv-highgui2.4v5, libopencv-imgproc2.4v5, libprotobuf9v5, libpython2.7 (>= 2.7), libstdc++6 (>= 5.2) Conflicts: caffe Filename: dists/xenial/main/binary-ppc64el/caffe-ibm_1.0.0rc3-3ibm4_ppc64el.deb Size: 14890854 MD5sum: 0a54fce6b68369a127a237562a62b562 SHA1: 544b5ce57a502dd081f69bc3198c3f2a51aa8c48 SHA256: 189c2cfbe3a1bbc3b1a2a3c5583121d80617f1d761a766d9c18ec06f02bdfa69 SHA512: fe4351ea1748311162ce1cf7d8481b88f7d0424df0539641e53c89db5d422f62d190433b308bdfb3dabff717d70bbdec01fa9afab618fcf0f788a23b0eb77ddd Homepage: http://caffe.berkeleyvision.org/ Description: Fast open framework for deep learning (BVLC upstream) Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and by community contributors. . This package includes BVLC's upstream Caffe with optimizations made by IBM. Package: python-engineio Architecture: all Version: 1.0.3-3ibm1 Priority: optional Section: python Maintainer: Frederic Bonnard Installed-Size: 71 Depends: python-six (>= 1.9.0), python:any (<< 2.8), python:any (>= 2.7.5-5~) Filename: dists/xenial/main/binary-ppc64el/python-engineio_1.0.3-3ibm1_all.deb Size: 15396 MD5sum: 18232696c3ccdb20612be598325d6016 SHA1: bdd34d5a219a146d151d093f3052feaec91992db SHA256: 721579cc5c42240b95633932cfcffb82d12c15a806a0469230628a0aa304c876 SHA512: 2e8b2fd0d9fa4c9a45bf8177d05e76b853c3fe269615c070a890add39e9b8f1248479bdb68a3c6d1a2315dd2463824a4dbe8e012d958b8adaca86222fa6872d6 Homepage: https://github.com/miguelgrinberg/python-engineio Description: Python implementation of the Engine.IO realtime server (Python 2) python-engineio project implements an Engine.IO server that can run standalone or integrated with a Python WSGI application. Engine.IO is a lightweight transport protocol that enables real-time bidirectional event-based communication between clients (typically web browsers) and a server. The official implementations of the client and server components are written in JavaScript. . This package contains the module for Python 2. Package: libopenmpi-dev-cuda Architecture: ppc64el Version: 2.0.1-4ibm1 Multi-Arch: foreign Priority: extra Section: libdevel Source: openmpi Maintainer: Alastair McKinstry Installed-Size: 4941 Depends: libc6 (>= 2.17), libopenmpi2-cuda (= 2.0.1-4ibm1), openmpi-common-cuda (= 2.0.1-4ibm1), libibverbs-dev (>= 1.1.7), libhwloc-dev Suggests: openmpi-doc-cuda Conflicts: libopenmpi-dev, libopenmpi-dev-cuda, libopenmpi2, libopenmpi2-cuda (<= 2.0.0-1), openmpi-bin, openmpi-bin-cuda (<= 1.2.4-0), openmpi-dev Breaks: libopenmpi1.10 (>= 1.10.2-1), libopenmpi1.6 (>= 1.6-1), libopenmpi2 (<= 2.0.0-1), libopenmpi2-cuda (<= 2.0.0-1) Replaces: libopenmpi1.10 (>= 1.10.2-1), libopenmpi1.6 (>= 1.6-1), libopenmpi2 (<= 2.0.0-1) Filename: dists/xenial/main/binary-ppc64el/libopenmpi-dev-cuda_2.0.1-4ibm1_ppc64el.deb Size: 1014578 MD5sum: 45d621f46dc44650004d099905b4051f SHA1: acfd146481d963e5e32ceafcd10ec6fc9a7c65b4 SHA256: bd97d5b44a2384b116b6e00a6fb405a3861b18d315db00d522b5513142178e37 SHA512: 04bb3909660f52034a37460e150b827a08068929bb1940e57a2a2c47e45099e35523ed336f7a09553b4a429a7fad8d0d88bc09550d13a60b70e1c526963013cc Homepage: http://www.open-mpi.org/ Description: high performance message passing library -- header files Open MPI is a project combining technologies and resources from several other projects (FT-MPI, LA-MPI, LAM/MPI, and PACX-MPI) in order to build the best MPI library available. A completely new MPI-3 compliant implementation, Open MPI offers advantages for system and software vendors, application developers and computer science researchers. . This package contains the header files and compiler wrappers which are needed to compile and link programs against libopenmpi. Package: libnccl-dev Architecture: ppc64el Version: 1.3.2-1.cuda8.0 Priority: optional Section: libdevel Source: nccl Maintainer: cudatools Installed-Size: 183 Depends: libnccl1 (= 1.3.2-1.cuda8.0) Filename: dists/xenial/main/binary-ppc64el/libnccl-dev_1.3.2-1.cuda8.0_ppc64el.deb Size: 149876 MD5sum: af05d7b95f6e641a8f2314d62dcb8be1 SHA1: c66a4dca46f8ff4d4179c70bf0a6748dca69e831 SHA256: 920fcb6b1d5fb37f0eb85c6e6e3aa44423f150cd40f9e9b96ee3e76359e63717 SHA512: 72f1a11a2384d156c1308ed88baf3a6606d5fe58ca89448b950c5615dac1103cf075f5d8d729e47a7d1dc7a8e467fb4d572793d08c393af3e6069bb7b18be229 Description: NVIDIA Collectives Communication Library (NCCL) Development Files NCCL (pronounced "Nickel") is a stand-alone library of standard collective communication routines for GPUs, such as all-gather, reduce, broadcast, etc., that have been optimized to achieve high bandwidth over PCIe. NCCL supports up to eight GPUs and can be used in either single- or multi-process (e.g., MPI) applications. Package: ddl-doc Architecture: ppc64el Version: 0.9.0-4ibm1 Priority: optional Section: doc Source: ddl Maintainer: IBM MLDL Installed-Size: 204 Filename: dists/xenial/main/binary-ppc64el/ddl-doc_0.9.0-4ibm1_ppc64el.deb Size: 180190 MD5sum: 6dc970dbfb8c2049e9b2ccbc2a9b12dd SHA1: 25ebf78e1317ae7a8d8ab55a6a990af9b2b3b3ea SHA256: cf739751ed034cbf7ffcd562ca5ede6a168d93832840c641307a43f7a152b96a SHA512: ce3b1f71cd719d192f6ba4b4428ac042cbc0a743256430c9a8074f2e462d9353b9b4f37160a9d9c2468dd56ea1c97ee4a71efeb545a978876cdbc9611185f45c Description: IBM PowerAI Distributed Deep Learning (DDL) This package contains documentation and examples for IBM PowerAI Distributed Deep Learning (DDL). Package: openmpi-doc-cuda Architecture: all Version: 2.0.1-4ibm1 Multi-Arch: foreign Priority: extra Section: doc Source: openmpi Maintainer: Alastair McKinstry Installed-Size: 1076 Conflicts: lam-mpidoc, lam4-dev, mpi-doc, openmpi-doc, openmpi-doc-cuda, openmpi-mpidoc Breaks: openmpi-checkpoint (<< 1.10.2) Replaces: openmpi-checkpoint (<< 1.10.2) Filename: dists/xenial/main/binary-ppc64el/openmpi-doc-cuda_2.0.1-4ibm1_all.deb Size: 703792 MD5sum: 619790ef7fad31d0e8f48d83df79083c SHA1: 43cb1a70c55ffd50bb1a1361fb2f5eb07a6343cb SHA256: 1141095f237067b0d85dfc4b106b4ec8949bfb88377c4dd9789eb74bc86993a1 SHA512: b97460795e6c10d24620567e058da821211ef8ed11324867c62a67e06d5d38eccaa05a1dc3abfd344aad6f503153399a3489c7189b2c499d9140caca0c5bbb75 Homepage: http://www.open-mpi.org/ Description: high performance message passing library -- man pages Open MPI is a project combining technologies and resources from several other projects (FT-MPI, LA-MPI, LAM/MPI, and PACX-MPI) in order to build the best MPI library available. A completely new MPI-3.1 compliant implementation, Open MPI offers advantages for system and software vendors, application developers and computer science researchers. . This package contains man pages describing the Message Passing Interface standard. Package: libopenblas Architecture: ppc64el Version: 0.2.19-3ibm2 Priority: optional Section: libs Maintainer: IBM MLDL Installed-Size: 36537 Depends: libgfortran3 (>= 4.3), libc6 (>= 2.17), libgomp1 (>= 4.9) Filename: dists/xenial/main/binary-ppc64el/libopenblas_0.2.19-3ibm2_ppc64el.deb Size: 4587040 MD5sum: 8ac5df06dce1be908ee32f5fd56c3683 SHA1: 1b77e9b8d8b01fac84d64eaac5dbfd3bfbc841dc SHA256: f9ed3b8d6a97eab1fb2d53e6604ce130c59932fe3b5859d4b0f8f49be8957c33 SHA512: f6e27e3bb26ef52605ae7887682f3c44487fca4957389df21ea1febedcfe0c0c08682bcbbd1031c8b505e0c3eaa0168319255b80f9332afae1ed245b010eb8c2 Homepage: http://www.openblas.net/ Description: Optimized BLAS library based on GotoBLAS2 1.13 BSD version More OpenBLAS infor available at: http://github.com/xianyi/OpenBLAS/wiki Package: tensorflow Architecture: ppc64el Version: 1.1.0-4ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 1275476 Depends: python (>= 2.7), python-mock (>= 1.3.0), python-numpy (>= 1.11.0), python-pip, python-setuptools, python-wheel, python-h5py, python-scipy, python-pandas, python-sklearn, swig, bazel, cuda-cublas-8-0, cuda-cudart-8-0, cuda-cufft-8-0, cuda-curand-8-0, cuda-license-8-0, libc6 (>= 2.23), libcudnn6, libgcc1 (>= 1:4.2), libstdc++6 (>= 5.2) Filename: dists/xenial/main/binary-ppc64el/tensorflow_1.1.0-4ibm1_ppc64el.deb Size: 120667226 MD5sum: 47c6106eec36c0af3c4a78e7f2738ef3 SHA1: 824d720a6f6ed5f6983a6ee00f44bcbe45a4ed55 SHA256: b6b0dfb9671032b193a46b82d77750afe12389653ca775693b1c829a1706b3aa SHA512: 366d3364e115b4b9c29cb70520a3db11480acc4ced1cd1dd9e30fa79a1c0b4292730e44e7707152b1137cd77b54fb4b77ad88dd5ae8fb6ba5eba022cd2851f88 Homepage: https://www.tensorflow.org/ Description: Open source software library for numerical computation TensorFlow is an open source software library for numerical computation using data flow graphs. Nodes in the graph represent mathematical operations, while the graph edges represent the multidimensional data arrays (tensors) communicated between them. Package: digits Architecture: ppc64el Version: 5.0-4ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 18587 Depends: build-essential, caffe-nv | caffe-bvlc | caffe-ibm, python-flask (>= 0.10.1), python-flask-socketio (>= 2.6.0), python-flaskext.wtf (>= 0.11), python-wtforms (>= 2.1), python-six (>= 1.5.2), python-requests (>= 2.2.1), python-setuptools (>= 3.3), python-eventlet (>= 0.13.0), python-pil (>= 2.3.0), python-numpy (>= 1.8.2), python-scipy (>= 0.13.3), python-protobuf (>= 2.5.0), python-lmdb (>= 0.87), python-gevent (>= 1.0), python-gevent-websocket (>= 0.9.3), python-h5py (>= 2.2.1), python-pydot, python-skimage, python-nose (>= 1.3.1), python-bs4 (>= 4.2.1), python-mock (>= 1.0.1), python-coverage (>= 3.7.1), python-selenium (>= 2.25.0), python-matplotlib (>= 1.3.1), python-psutil (>= 3.4.2), python-scikit-fmm (>= 0.0.9), python (>= 2.7), libhdf5-serial-dev Recommends: torch Conflicts: python-socketio Filename: dists/xenial/main/binary-ppc64el/digits_5.0-4ibm1_ppc64el.deb Size: 10809198 MD5sum: 13a9ee7e6024d0f6c311bba6f785903a SHA1: bdbf558c128184e0b581e1462ca9eb05c2ce988d SHA256: 99f0b7188db38e7472dd98d7c974803d9d7f40f6466b1969b9914a1d8586dacc SHA512: cef0a2f4e382152be4469f3f827b53f59cd4dce00afa7b340782017863d9119b6f6b845408cd42c01462eb09267c4c84c8c8451e2ed19748c7f9904a4dbe9e50 Homepage: https://developer.nvidia.com/digits Description: Deep Learning GPU Training System DIGITS (the Deep Learning GPU Training System) is a webapp for training deep learning models. Package: libopenblas Architecture: ppc64el Version: 0.2.19-4ibm1 Priority: optional Section: libs Maintainer: IBM MLDL Installed-Size: 36534 Depends: libgfortran3 (>= 4.3), libc6 (>= 2.17), libgomp1 (>= 4.9) Filename: dists/xenial/main/binary-ppc64el/libopenblas_0.2.19-4ibm1_ppc64el.deb Size: 4586246 MD5sum: 64b3681d06ee2368727c62172445840d SHA1: f51a21f809e9c6d43454592ee45bd5df5bda1323 SHA256: 58e9c1dabbd7e3ffe278413271222f81c017c4a2382b26df1c76edc9de0994f2 SHA512: e84d594f00f5bd86dede5108c3c283cb543998f447e7efa4c5b446cd5f20fe148e7fa7bea2e57e46d42f500449cc8891fed55357c241edc9b49a457f2c77fd1a Homepage: http://www.openblas.net/ Description: Optimized BLAS library based on GotoBLAS2 1.13 BSD version More OpenBLAS infor available at: http://github.com/xianyi/OpenBLAS/wiki Package: openmpi-common-cuda Architecture: all Version: 2.0.1-4ibm1 Multi-Arch: foreign Priority: extra Section: net Source: openmpi Maintainer: Alastair McKinstry Installed-Size: 395 Conflicts: openmpi-common, openmpi-common-cuda Filename: dists/xenial/main/binary-ppc64el/openmpi-common-cuda_2.0.1-4ibm1_all.deb Size: 149722 MD5sum: d2abd65c0ccbe87fb700d07d8da46052 SHA1: d289337f0a9af1dfabcb9b92bd4f9f5e1ca17327 SHA256: 2561e732d0ee05387ec3e587cf5f3c5e7e5512e872f4fd1cd605e164c6c2e0b3 SHA512: f86ae5561254252fd8140f4eb2d5458917280bb08d8b8d191e1f413881bce55dc76428fe2fc45200f9046089c9a415889bad3165e7952f7073401de77f87be2d Homepage: http://www.open-mpi.org/ Description: high performance message passing library -- common files Open MPI is a project combining technologies and resources from several other projects (FT-MPI, LA-MPI, LAM/MPI, and PACX-MPI) in order to build the best MPI library available. A completely new MPI-3.1 compliant implementation, Open MPI offers advantages for system and software vendors, application developers and computer science researchers. . This package contains platform independent files for Open MPI. Package: ddl-tensorflow Architecture: ppc64el Version: 0.9.0-4ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 2471 Depends: cuda-cublas-8-0, libc6 (>= 2.17), libgcc1 (>= 1:3.0), libgoogle-glog0v5, libnccl1, libopenmpi2-cuda, libstdc++6 (>= 5.2), ddl-doc, tensorflow Filename: dists/xenial/main/binary-ppc64el/ddl-tensorflow_0.9.0-4ibm1_ppc64el.deb Size: 1364728 MD5sum: 1ab622ce4a10cb9cec1a8f123c988cf9 SHA1: bdaf1f26f4ea562b2daf14ee8b1e1725f7923cb8 SHA256: 2d00a399089a671b4942d5df5fddbb9d6aebbf22571d2ea06c7fdd092331b89a SHA512: 230e257eb0c923e316475c17c398bba44f50816dd9dd26fae8e6eaad330f5eeeb0dfce999fc1aabd41feadd32dcf96cddf362126a5dcc3d1d8eec51e2a3533f5 Description: TensorFlow operator for IBM PowerAI Distributed Deep Learning Package: caffe-bvlc Architecture: ppc64el Version: 1.0.0rc5-3ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 94275 Depends: build-essential, cmake, curl, g++, gcc, git-core, protobuf-compiler, python (>= 2.7), python-skimage, python-protobuf, libnccl1, cuda-cublas-8-0, cuda-cudart-8-0, cuda-curand-8-0, libboost-filesystem1.58.0, libboost-python1.58.0, libboost-system1.58.0, libboost-thread1.58.0, libc6 (>= 2.17), libcudnn5-dev, libgcc1 (>= 1:3.0), libgflags2v5, libgoogle-glog0v5, libhdf5-10, libleveldb1v5, liblmdb0 (>= 0.9.7), libopenblas, libopencv-core2.4v5, libopencv-highgui2.4v5, libopencv-imgproc2.4v5, libprotobuf9v5, libpython2.7 (>= 2.7), libstdc++6 (>= 5.2) Conflicts: caffe Filename: dists/xenial/main/binary-ppc64el/caffe-bvlc_1.0.0rc5-3ibm1_ppc64el.deb Size: 15104722 MD5sum: 17ba365727ddc4b1dd561949567a946e SHA1: 844c64ce35d25e49a8d9bad6b163defc5f3ca4de SHA256: 272bb25b58fa0690c15d5100f575e4b467aa5e90dd6939749cb68052f99c758c SHA512: 72ca77bf1afcd0fed05e8746aca82979f921db524f7232a84f3408a0ed296241b6a508fbd45889d842af64dbe2118d54126fa29d4bc6c38368fab6b15c136891 Homepage: http://caffe.berkeleyvision.org/ Description: Fast open framework for deep learning (BVLC upstream) Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and by community contributors. . This package includes BVLC's upstream Caffe. Package: python3-engineio Architecture: all Version: 1.0.3-3ibm1 Priority: optional Section: python Source: python-engineio Maintainer: Frederic Bonnard Installed-Size: 71 Depends: python3-six (>= 1.9.0), python3:any (>= 3.3.2-2~) Filename: dists/xenial/main/binary-ppc64el/python3-engineio_1.0.3-3ibm1_all.deb Size: 15482 MD5sum: 07bc59e96e9463992b4f456f7a5c7d5f SHA1: 8bd218e39fc2f2216b510b7f2477e78ee8925fed SHA256: 9c243243b62f7d9204b134f73745a1820335b35181199dbc8427f62a2a6c0ed1 SHA512: d5deac78ed0cc8a0f18dffe67c5501b39d35d631e58b9fd3cb55a14d88efd025616d3a2c0c27f49a34bd29606ebed3c6459fb562a777218646ac983b486c2c50 Homepage: https://github.com/miguelgrinberg/python-engineio Description: Python implementation of the Engine.IO realtime server (Python 3) python-engineio project implements an Engine.IO server that can run standalone or integrated with a Python WSGI application. Engine.IO is a lightweight transport protocol that enables real-time bidirectional event-based communication between clients (typically web browsers) and a server. The official implementations of the client and server components are written in JavaScript. . This package contains the module for Python 3. Package: tensorflow Architecture: ppc64el Version: 1.0.1-3ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 985343 Depends: python (>= 2.7), python-mock (>= 1.3.0), python-numpy (>= 1.11.0), python-pip, python-setuptools, python-wheel, swig, bazel, cuda-cudart-8-0, libc6 (>= 2.23), libgcc1 (>= 1:4.2), libstdc++6 (>= 5.2) Filename: dists/xenial/main/binary-ppc64el/tensorflow_1.0.1-3ibm1_ppc64el.deb Size: 91031258 MD5sum: d7b2d92d73d8d2afd39decf299ed968c SHA1: 0f5446ccac603793daa1db3f365d058d01d9e832 SHA256: 425e0949507e00d09bdd905bcf506bc81e7953d4bed858a856c918c5c73fc5e7 SHA512: 833d897f36ca3534a39f531154f38951d6214765bc916cf1f40613c2707f8e9d99b08f1f313c22a337532ae755c90c1f6b1fb8b9069aa69d2ca2641942530fa7 Homepage: https://www.tensorflow.org/ Description: Open source software library for numerical computation TensorFlow is an open source software library for numerical computation using data flow graphs. Nodes in the graph represent mathematical operations, while the graph edges represent the multidimensional data arrays (tensors) communicated between them. Package: caffe-nv Architecture: ppc64el Version: 0.15.14-4ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 93855 Depends: build-essential, g++, gcc, protobuf-compiler, python (>= 2.7), python-skimage, python-protobuf, libnccl1, cuda-cublas-8-0, cuda-cudart-8-0, cuda-curand-8-0, libboost-filesystem1.58.0, libboost-python1.58.0, libboost-system1.58.0, libboost-thread1.58.0, libc6 (>= 2.17), libcudnn6, libgcc1 (>= 1:3.0), libgflags2v5, libgoogle-glog0v5, libhdf5-10, libleveldb1v5, liblmdb0 (>= 0.9.7), libopenblas, libopencv-core2.4v5, libopencv-highgui2.4v5, libopencv-imgproc2.4v5, libprotobuf9v5, libpython2.7 (>= 2.7), libstdc++6 (>= 5.2) Conflicts: caffe Filename: dists/xenial/main/binary-ppc64el/caffe-nv_0.15.14-4ibm1_ppc64el.deb Size: 15212874 MD5sum: 99969f08d9c9077aea4ba672727a5271 SHA1: c465309051d1b2490bb60673fcf2e804b596a52d SHA256: 2326f2ceab49e6c03ea932c4efe4456b2955989d8f0dd1e6037451dc2ba2fa6b SHA512: 6a07d72442488a72b42b165c913183f09dd111314e87bfc2c26323e22c7eb7d2ea3913f6b7bc6851fcbc38f8ed7b9225b8d8a21082b9626adb752f87ba9f363f Homepage: http://caffe.berkeleyvision.org/ Description: Fast open framework for deep learning (NVIDIA's fork) Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and by community contributors. . This package includes NVIDIA's fork of upstream Caffe. Package: openmpi-bin-cuda Architecture: ppc64el Version: 2.0.1-4ibm1 Priority: extra Section: net Source: openmpi Maintainer: Alastair McKinstry Installed-Size: 882 Depends: libc6 (>= 2.17), libopenmpi2-cuda, openmpi-common-cuda (= 2.0.1-4ibm1) Suggests: gfortran Conflicts: openmpi-bin, openmpi-bin-cuda Filename: dists/xenial/main/binary-ppc64el/openmpi-bin-cuda_2.0.1-4ibm1_ppc64el.deb Size: 183118 MD5sum: 53f76ebc256c3990a4cdf22dafe26289 SHA1: c189e38bf85ad72420f2d9605e9ea83d734af8ba SHA256: 15b9dfcc3c86a7da370df176b24239c119a89bfe402dd0c379316600e5aac0c9 SHA512: 8b5bd894614ac7f126e5e7a32612d5e58a3f685baa369b0cb8bc6a5017685399d2d84800800e4f39919fa6255659874cc5276dfe052c5270bccfb8db6ace4745 Homepage: http://www.open-mpi.org/ Description: high performance message passing library -- binaries Open MPI is a project combining technologies and resources from several other projects (FT-MPI, LA-MPI, LAM/MPI, and PACX-MPI) in order to build the best MPI library available. A completely new MPI-3.1 compliant implementation, Open MPI offers advantages for system and software vendors, application developers and computer science researchers. . Features: * Full MPI-3.1 standards conformance * Thread safety and concurrency * Dynamic process spawning * High performance on all platforms * Reliable and fast job management * Network and process fault tolerance * Support network heterogeneity * Single library supports all networks * Run-time instrumentation * Many job schedulers supported * Internationalized error messages * Component-based design, documented APIs . This package contains the Open MPI utility programs. Package: python-lmdb Architecture: ppc64el Version: 0.87-3ibm1 Priority: optional Section: python Source: py-lmdb Maintainer: Ubuntu Developers Original-Maintainer: Debian Berkeley DB Group Installed-Size: 260 Depends: libc6 (>= 2.17), liblmdb0 (>= 0.9.14), python (<< 2.8), python (>= 2.7~), python:any (>= 2.7.5-5~) Filename: dists/xenial/main/binary-ppc64el/python-lmdb_0.87-3ibm1_ppc64el.deb Size: 44140 MD5sum: 99636255352a17d7196aade3a943c002 SHA1: 5f6627b5df7307913007f9a5e76fc24d2c0091ce SHA256: fcb6658638cb0deca45d3de428d46926027affea21acc1fadccd6e516905ceb3 SHA512: 1d2e256ac5acf22b26e7a3e2dcce3444683a369c3499465ed33e1eb1c24748c57575bedccba52a341a46e7b15ff487aa35019f85469ff0a73b76639ee61f0587 Homepage: https://github.com/dw/py-lmdb Description: Python binding for LMDB Lightning Memory-Mapped Database Lighting Memory-Mapped Database (LMDB) is an ultra-fast, ultra-compact key-value embedded data store developed for the OpenLDAP Project. It uses memory-mapped files, so it has the read performance of a pure in-memory database while still offering the persistence of standard disk-based databases, and is only limited to the size of the virtual address space (it is not limited to the size of physical RAM). . This package contains the 'lmdb' Python extension module. Package: python3-socketio-server Architecture: all Version: 1.6.0-3ibm1 Priority: optional Section: python Source: python-socketio-server Maintainer: Frederic Bonnard Installed-Size: 79 Depends: python3-engineio, python3-six (>= 1.9.0), python3:any (>= 3.3.2-2~) Filename: dists/xenial/main/binary-ppc64el/python3-socketio-server_1.6.0-3ibm1_all.deb Size: 16506 MD5sum: 807e98030dc3251681650acd7c52edae SHA1: 5b9e615cc044bfeeef82754c10ce2676c637146c SHA256: f773596c7e324f192ae41262fd7eb2ce73173443e11be65550ecccdb3ceef38f SHA512: c0f0de9678e065d46ec5552f89abf955f1901877e6bc2fb1b1aab43c8a51ebd57b842b4894accc3859efdf7652029a54b143566060f3b83a8de1f6c4381dfe09 Homepage: https://github.com/miguelgrinberg/python-socketio Description: Python implementation of the Socket.IO realtime server (Python 3) python-engineio project implements a Socket.IO server that can run standalone or integrated with a Python WSGI application. Socket.IO is a transport protocol that enables real-time bidirectional event-based communication between clients (typically web browsers) and a server. The original implementations of the client and server components are written in JavaScript. . This package contains the module for Python 3. Package: tensorflow Architecture: ppc64el Version: 0.12.0-3ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 377199 Depends: python (>= 2.7), python-mock (>= 1.3.0), python-numpy (>= 1.11.0), python-pip, python-setuptools, python-wheel, swig, bazel, cuda-cudart-8-0, libc6 (>= 2.23), libgcc1 (>= 1:3.0), libstdc++6 (>= 5.2) Filename: dists/xenial/main/binary-ppc64el/tensorflow_0.12.0-3ibm1_ppc64el.deb Size: 33492440 MD5sum: 3ad36cc644fdc77310c77209e128530c SHA1: a504cba1b71f9256695612e2f35c9e7f30bbd373 SHA256: c8be4d76b2b31b6994db357ddc67756da53023594874e7e3b23346c4ae7af152 SHA512: 82e63f20a93a3ead1e6ab491de0f579e13261bbe2a90812984e438966c13e6c1691a16a220174a6c3ef332b4305a170043f4a0640eed9c53fa65252644991cb8 Homepage: https://www.tensorflow.org/ Description: Open source software library for numerical computation TensorFlow is an open source software library for numerical computation using data flow graphs. Nodes in the graph represent mathematical operations, while the graph edges represent the multidimensional data arrays (tensors) communicated between them. Package: digits Architecture: ppc64el Version: 5.0-3ibm5 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 18585 Depends: build-essential, caffe-nv | caffe-bvlc | caffe-ibm, python-flask (>= 0.10.1), python-flask-socketio (>= 2.6.0), python-flaskext.wtf (>= 0.11), python-wtforms (>= 2.1), python-six (>= 1.5.2), python-requests (>= 2.2.1), python-setuptools (>= 3.3), python-eventlet (>= 0.13.0), python-pil (>= 2.3.0), python-numpy (>= 1.8.2), python-scipy (>= 0.13.3), python-protobuf (>= 2.5.0), python-lmdb (>= 0.87), python-gevent (>= 1.0), python-gevent-websocket (>= 0.9.3), python-h5py (>= 2.2.1), python-pydot, python-skimage, python-nose (>= 1.3.1), python-bs4 (>= 4.2.1), python-mock (>= 1.0.1), python-coverage (>= 3.7.1), python-selenium (>= 2.25.0), python-matplotlib (>= 1.3.1), python-psutil (>= 3.4.2), python-scikit-fmm (>= 0.0.9), python (>= 2.7), libhdf5-serial-dev Recommends: torch Conflicts: python-socketio Filename: dists/xenial/main/binary-ppc64el/digits_5.0-3ibm5_ppc64el.deb Size: 10810726 MD5sum: 0f010711da1bfe340db14482cc8bf67d SHA1: d4b9194133df993d369431b0bd6641915865eacf SHA256: 4b34443bda3a0acde5934a224a8da79c8a772ea803e2b9c5c75dd9d013caba18 SHA512: 924ead4457ed2ad7ffe4ef1ad0923a1ab1261cb4435a8ccb0ef4c7a8975b6f2b4a660e8af36b51b75f2567b614fe394ddf3de023f0599a1f6300e0c6774bca31 Homepage: https://developer.nvidia.com/digits Description: Deep Learning GPU Training System DIGITS (the Deep Learning GPU Training System) is a webapp for training deep learning models. Package: libopenmpi2-cuda Architecture: ppc64el Version: 2.0.1-4ibm1 Multi-Arch: same Priority: extra Section: libs Source: openmpi Maintainer: Alastair McKinstry Installed-Size: 8931 Depends: libc6 (>= 2.17), libgcc1 (>= 1:3.0), libhwloc5 (>= 1.11.2), libibverbs1 (>= 1.1.2), libstdc++6 (>= 4.1.1), libhwloc-plugins Recommends: openmpi-bin-cuda Conflicts: libopenmpi1.10, libopenmpi1.6, libopenmpi2 Breaks: libopenmpi1.10 (>= 1.10.2-1) Replaces: libopenmpi1.10 (>= 1.10.2-1) Filename: dists/xenial/main/binary-ppc64el/libopenmpi2-cuda_2.0.1-4ibm1_ppc64el.deb Size: 2039624 MD5sum: 03977d09a25597e632b11c49ae734205 SHA1: 1f90b252ca1347194ae04a49fda11f366a6ed2d0 SHA256: dda2a2f912b375722f3b36b0544781262f5263ed95fbe3b1c50f8ce33851e6db SHA512: b10ba05dc9f0de1f4dbf9426c3940e125ef695c54c498e3beb1e1fc5632356b655bf0365c8f1950f08a7abccf272eff71f99ffd382256817e5795b77980a75cf Homepage: http://www.open-mpi.org/ Description: high performance message passing library -- shared library Open MPI is a project combining technologies and resources from several other projects (FT-MPI, LA-MPI, LAM/MPI, and PACX-MPI) in order to build the best MPI library available. A completely new MPI-3.1 compliant implementation, Open MPI offers advantages for system and software vendors, application developers and computer science researchers. . This package contains the Open MPI shared libraries. Package: libnccl-dev Architecture: ppc64el Version: 1.3.2-4ibm3.cuda8.0 Priority: optional Section: libdevel Source: nccl Maintainer: cudatools Installed-Size: 183 Depends: libnccl1 (= 1.3.2-4ibm3.cuda8.0) Filename: dists/xenial/main/binary-ppc64el/libnccl-dev_1.3.2-4ibm3.cuda8.0_ppc64el.deb Size: 149922 MD5sum: cf631c3dbd5218385921f29b34bd1d4a SHA1: 405458318ee4fda09e4678b366b2ade7e17cd2df SHA256: fb49d5cf5c38d1670d2fb480ee4d43afea292eacfa51cbf26ca3150781123046 SHA512: 325b6ff145e91b454d2d3c680732fdc0c479c68b525812db62ad552c4d849afd0352bceb5c1914abdc86da214c8478206e464a416416b26a1a3ef6d7ecd6a083 Description: NVIDIA Collectives Communication Library (NCCL) Development Files NCCL (pronounced "Nickel") is a stand-alone library of standard collective communication routines for GPUs, such as all-gather, reduce, broadcast, etc., that have been optimized to achieve high bandwidth over PCIe. NCCL supports up to eight GPUs and can be used in either single- or multi-process (e.g., MPI) applications. Package: python3-lmdb Architecture: ppc64el Version: 0.87-3ibm1 Priority: optional Section: python Source: py-lmdb Maintainer: Ubuntu Developers Original-Maintainer: Debian Berkeley DB Group Installed-Size: 260 Depends: libc6 (>= 2.17), liblmdb0 (>= 0.9.14), python3 (<< 3.6), python3 (>= 3.5~) Filename: dists/xenial/main/binary-ppc64el/python3-lmdb_0.87-3ibm1_ppc64el.deb Size: 44664 MD5sum: 465bc38fec33ab7861e32cc55b2f7151 SHA1: 65dc4936351c6d7243e91b4c62c87377fde95202 SHA256: 26abceb030bcba7ec03113918d3216a57c8f0c74b994c581d41a9e18534fd3ad SHA512: 41e2274f3b82895b509f1304debb068d26670b7646c3b6646ade44706079418b484ce3b203a7565c962cd68dc45bb1e5c2b426a502473c994a7032430fe5db9d Homepage: https://github.com/dw/py-lmdb Description: Python 3 binding for LMDB Lightning Memory-Mapped Database Lighting Memory-Mapped Database (LMDB) is an ultra-fast, ultra-compact key-value embedded data store developed for the OpenLDAP Project. It uses memory-mapped files, so it has the read performance of a pure in-memory database while still offering the persistence of standard disk-based databases, and is only limited to the size of the virtual address space (it is not limited to the size of physical RAM). . This package contains the 'lmdb' Python 3 extension module. Package: chainer Architecture: ppc64el Version: 1.20.0.1-3ibm1 Priority: optional Section: python Maintainer: IBM MLDL Installed-Size: 26315 Depends: python (>= 2.7.6), python-numpy (>= 1.9.0), python-six (>= 1.9.0), python-nose, python-h5py (>= 2.5.0), python-mock, python-protobuf, libcudnn5, cuda-cublas-8-0, cuda-cudart-8-0, cuda-curand-8-0, libc6 (>= 2.17), libfreetype6 (>= 2.2.1), libgcc1 (>= 1:3.0), libjpeg8 (>= 8c), liblcms2-2 (>= 2.2+git20110628), libstdc++6 (>= 4.1.1), libtiff5 (>= 4.0.3), zlib1g (>= 1:1.1.4) Filename: dists/xenial/main/binary-ppc64el/chainer_1.20.0.1-3ibm1_ppc64el.deb Size: 10094546 MD5sum: f540a6f079597e085fd0d568be056dba SHA1: 4502c2531758f8e19edf25c60cd7982922acc833 SHA256: 31c6018c4db72ed988abda46b27a05e58d6140d55f384e83a610a780b3e534c9 SHA512: c3d8151e2eea3c9e09a75fcce21287304d364c72680edd0ec32f04b1094751be2a6c610a2519d5c89b3f0629feca310ce7899cd0416791c3d7051ac5726a0977 Homepage: http://chainer.org/ Description: A flexible framework of neural networks for deep learning A flexible framework of neural networks for deep learning, including GPU support. Package: theano Architecture: ppc64el Version: 0.9.0-4ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 35302 Depends: libopenblas, python (>= 2.7), python-dev, python-nose, python-numpy (>= 1.9.1), python-scipy (>= 0.14), python-mako (>= 0.7), python-six, g++, libc6 (>= 2.17) Recommends: python-flake8, python-pydot Filename: dists/xenial/main/binary-ppc64el/theano_0.9.0-4ibm1_ppc64el.deb Size: 9762464 MD5sum: 4b09387e9f0b98ee2b4014f5406934e7 SHA1: e6f6f5c47a5e68a11667b5bdfa2f1e0f6e59005f SHA256: 652b0fb597c9eb01be49d799ceb9702676e95d32e7b51d74bf7b76b9041db369 SHA512: 6a92cb8e114e3ad462ae2530fb1c08041287f62f1262ada911c7757ce647db03c4dbc6982a78af8c7df8af5b5c18fe6323a568f97740367f0233f757b6ce4d9c Homepage: http://www.deeplearning.net/software/theano/ Description: A Python library for Deep Learning Theano is a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently. This package includes support for NVIDIA GPUs, and requires NVIDIA CUDA Toolkit and CuDNN. Package: caffe-ibm Architecture: ppc64el Version: 1.0.0.-4ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 103270 Depends: build-essential, ddl-doc, g++, gcc, libnccl1, libopenmpi2-cuda, openmpi-bin-cuda, openmpi-common-cuda, openmpi-doc-cuda, protobuf-compiler, python (>= 2.7), python-protobuf, python-skimage, cuda-cublas-8-0, cuda-cudart-8-0, cuda-curand-8-0, libboost-filesystem1.58.0, libboost-python1.58.0, libboost-system1.58.0, libboost-thread1.58.0, libc6 (>= 2.17), libcudnn6, libgcc1 (>= 1:3.0), libgflags2v5, libgomp1 (>= 4.9), libgoogle-glog0v5, libhdf5-10, libleveldb1v5, liblmdb0 (>= 0.9.7), libopenblas, libopencv-core2.4v5, libopencv-highgui2.4v5, libopencv-imgproc2.4v5, libprotobuf9v5, libpython2.7 (>= 2.7), libstdc++6 (>= 5.2) Conflicts: caffe Filename: dists/xenial/main/binary-ppc64el/caffe-ibm_1.0.0.-4ibm1_ppc64el.deb Size: 15812572 MD5sum: fb9e7892cbc793072325874d939672a8 SHA1: 6526d69572cace7802c191e3b3cf75cb9f14f4fb SHA256: f4e4c005e65a911e59cf87dde096c1fb5ac984b66be0ef642a5dfe7e2b4eb032 SHA512: 4d72ae3a2c48f30c4f5cfb7514d3e193fc59df309ca58fac6bdd7e15d0ffc7cbcc98164d799655d2e14691c58d057d2b8c722abca23837fc4625155c002ed822 Homepage: http://caffe.berkeleyvision.org/ Description: Fast open framework for deep learning (BVLC upstream) Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and by community contributors. . This package includes BVLC's upstream Caffe with optimizations made by IBM. Package: power-mldl Architecture: ppc64el Version: 4.0.0 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 4 Depends: bazel (>= 0.4.5-4ibm1), caffe-bvlc (>= 1.0.0.-4ibm1), caffe-ibm (>= 1.0.0.-4ibm1), caffe-nv (>= 0.15.14-4ibm1), chainer (>= 1.23.0-4ibm1), digits (>= 5.0-4ibm1), libopenblas (>= 0.2.19-4ibm1), libnccl1 (>= 1.3.2-4ibm1.cuda8.0), libnccl-dev (>= 1.3.2-4ibm1.cuda8.0), tensorflow (>= 1.1.0-4ibm1), ddl-tensorflow (>= 0.9.0-4ibm1), theano (>= 0.9.0-4ibm1), torch (>= 7-4ibm1) Filename: dists/xenial/main/binary-ppc64el/power-mldl_4.0.0_ppc64el.deb Size: 2318 MD5sum: d6801794edf38b59c642f3131d7fb59e SHA1: a8b4c6e71e5a09facba54dfcbabdc889226c9b11 SHA256: 7429707b99534904ddc0762aa67ef95243fe57eaa350e729e4ff6516b0b03d87 SHA512: 0a3befdbd44610deeb49ab583c86c8e5bd46e63e7f24e4490a882165a094c927117d71af07d876ca3bbf0e9951f7625985fea85f4ce399ea717a8b2151115cce Description: Meta-package for Deep Learning frameworks for POWER This package does not include any Deep Learning binaries. It is a convenience to allow easy install of the Deep Learning frameworks currently packaged by IBM. Package: chainer Architecture: ppc64el Version: 1.23.0-4ibm1 Priority: optional Section: python Maintainer: IBM MLDL Installed-Size: 27097 Depends: python (>= 2.7.6), python-numexpr, python-numpy (>= 1.9.0), python-six (>= 1.9.0), python-nose, python-h5py (>= 2.5.0), python-mock, python-protobuf, libcudnn6, libopenblas, cuda-cublas-8-0, cuda-cudart-8-0, cuda-curand-8-0, cuda-cusolver-8-0, cuda-license-8-0, libc6 (>= 2.17), libfreetype6 (>= 2.2.1), libgcc1 (>= 1:3.0), libjpeg8 (>= 8c), liblcms2-2 (>= 2.2+git20110628), libstdc++6 (>= 4.1.1), libtiff5 (>= 4.0.3), zlib1g (>= 1:1.1.4) Filename: dists/xenial/main/binary-ppc64el/chainer_1.23.0-4ibm1_ppc64el.deb Size: 10193490 MD5sum: 1228093155f968000c1a031fad4eac77 SHA1: 2464bd3cec66bf665548ef9814de6e8cd3c77ccd SHA256: d6b77ea7e06d82983177e7795bba1d835b36bcc79602ff099205716080ca9fbf SHA512: 703e8562dc5cb0fe1c0b94221d1d525d2677b665cd55b5ef5b665af323f754270840269965ad8c0415f51aec29470f99e5d36c421ee38c622e215daa7e2b9852 Homepage: http://chainer.org/ Description: A flexible framework of neural networks for deep learning A flexible framework of neural networks for deep learning, including GPU support. Package: caffe-bvlc Architecture: ppc64el Version: 1.0.0.-4ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 100637 Depends: build-essential, g++, gcc, protobuf-compiler, python (>= 2.7), python-skimage, python-protobuf, libnccl1, cuda-cublas-8-0, cuda-cudart-8-0, cuda-curand-8-0, libboost-filesystem1.58.0, libboost-python1.58.0, libboost-system1.58.0, libboost-thread1.58.0, libc6 (>= 2.17), libcudnn6, libgcc1 (>= 1:3.0), libgflags2v5, libgoogle-glog0v5, libhdf5-10, libleveldb1v5, liblmdb0 (>= 0.9.7), libopenblas, libopencv-core2.4v5, libopencv-highgui2.4v5, libopencv-imgproc2.4v5, libprotobuf9v5, libpython2.7 (>= 2.7), libstdc++6 (>= 5.2) Conflicts: caffe Filename: dists/xenial/main/binary-ppc64el/caffe-bvlc_1.0.0.-4ibm1_ppc64el.deb Size: 15272484 MD5sum: cd5ae4a0586daefb699f21cc3b0e26bb SHA1: e3249696024f8291a6750262719b753a31c053ec SHA256: 0825435439bcc677948eb87f379a1b615d539ddb95e1f8c0a87960dcae4c0822 SHA512: 029767c138fcd28713ed734683ac488c0ffd12a851f06e22ff513d180a20ecd66964c872219646d04a82845e0d9320bf00dcbb59a76d44d4905932a69a882f50 Homepage: http://caffe.berkeleyvision.org/ Description: Fast open framework for deep learning (BVLC upstream) Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and by community contributors. . This package includes BVLC's upstream Caffe. Package: python-scikit-fmm Architecture: ppc64el Version: 0.0.9-3ibm1 Priority: optional Section: python Source: scikit-fmm Maintainer: Frederic Bonnard Installed-Size: 213 Depends: python-numpy (>= 1:1.10.0~b1), python-numpy-abi9, python (<< 2.8), python (>= 2.7~), python:any (>= 2.7.5-5~), libc6 (>= 2.17), libgcc1 (>= 1:3.0), libstdc++6 (>= 5.2) Suggests: python-scikit-fmm-doc Filename: dists/xenial/main/binary-ppc64el/python-scikit-fmm_0.0.9-3ibm1_ppc64el.deb Size: 53272 MD5sum: 0f7025b461f4212d60623b0a13715f24 SHA1: aaf2359b1070f8faff04545d87d02da181e4edcb SHA256: dccf5e61219f1ca4c9903f0ecdde1a7ce9c4e864a4e4723ee9ae30ea313a96b2 SHA512: 4c3823439d3f3551018bc9cc3eac6b793eff29f814392f7c488c57b25bba60dbcb4a03ebf461ebcc86b98822a8f6dbbc300156c5fb30a13ca2c64469eae82481 Homepage: https://github.com/scikit-fmm/scikit-fmm Description: Fast marching method for Python (Python 2) scikit-fmm is a Python extension module which implements the fast marching method. The fast marching method is used to model the evolution of boundaries and interfaces in a variety of application areas. More specifically, the fast marching method is a numerical technique for finding approximate solutions to boundary value problems of the Eikonal equation. . This package contains the module for Python 2. Package: python-scikit-fmm-doc Architecture: all Version: 0.0.9-3ibm1 Priority: optional Section: doc Source: scikit-fmm Maintainer: Frederic Bonnard Installed-Size: 856 Depends: libjs-sphinxdoc (>= 1.0) Filename: dists/xenial/main/binary-ppc64el/python-scikit-fmm-doc_0.0.9-3ibm1_all.deb Size: 398526 MD5sum: e1a12657a37fb93e963833ca6bbdf358 SHA1: d25d174965cc909d62d4b47a4c7d3869becaf251 SHA256: ce5c42a4c8d165d33746ebb0af8b658c38f47ca9f90ba43e1119a60883e6e608 SHA512: 53c9e3736f2daabc6c65eb2fea275cbb831f700225985549f50b09f03ed14dcb2a5c6161447a831b373335ba643fa1915cdde2d3fe3ad0161f72e023bd9fb5d2 Homepage: https://github.com/scikit-fmm/scikit-fmm Description: Fast marching method for Python (documentation) scikit-fmm is a Python extension module which implements the fast marching method. The fast marching method is used to model the evolution of boundaries and interfaces in a variety of application areas. More specifically, the fast marching method is a numerical technique for finding approximate solutions to boundary value problems of the Eikonal equation. . This package contains the documentation. Package: libnccl1 Architecture: ppc64el Version: 1.3.2-4ibm3.cuda8.0 Priority: optional Section: libs Source: nccl Maintainer: cudatools Installed-Size: 50985 Depends: cuda-cudart-8-0, libc6 (>= 2.17), libgcc1 (>= 1:3.0), libstdc++6 (>= 4.1.1) Filename: dists/xenial/main/binary-ppc64el/libnccl1_1.3.2-4ibm3.cuda8.0_ppc64el.deb Size: 1776638 MD5sum: 9b871785f0b653d9221c29b7e22a149e SHA1: 2d1ce5a38c0669ba689aac205852177e77f10412 SHA256: 09f63a1451f294ccdebac5b49805b2c38e512900d328d266ae9415511a139163 SHA512: 7760ae26da3b99827727298fb2e8bf4e3293647c06fb1df3764f8f3132cb26bb9b0d32998d97ac56143f4e21242fc19738f3b9d1eb1280e90b0079ddc12eebb8 Description: NVIDIA Collectives Communication Library (NCCL) Runtime NCCL (pronounced "Nickel") is a stand-alone library of standard collective communication routines for GPUs, such as all-gather, reduce, broadcast, etc., that have been optimized to achieve high bandwidth over PCIe. NCCL supports up to eight GPUs and can be used in either single- or multi-process (e.g., MPI) applications. Package: bazel Architecture: ppc64el Version: 0.4.5-4ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 130319 Depends: libc6 (>= 2.17), libgcc1 (>= 1:3.0), libstdc++6 (>= 5.2) Filename: dists/xenial/main/binary-ppc64el/bazel_0.4.5-4ibm1_ppc64el.deb Size: 128588658 MD5sum: 9bda248291f31455574d880c1d03ab34 SHA1: 28c77d4cee44d8e4d2793c567883f179ee82316b SHA256: 52ad5cd91e28f8bb98ed0cc496ad0bfff08999fc731629484a754ee585b22552 SHA512: 43ca7d0b7a0a5bc8c2a225bfe763cc2a4f14da8b26074d650a0d5827afda86766f4a55a0debd65a1200b3c3506faf599e38a0ce77e1defc7eedc72141bcdefdc Homepage: http://www.bazel.io/ Description: Correct, reproducible, fast builds for everyone Bazel is Google's own build tool, now publicly available in Beta. Bazel has built-in support for building both client and server software, including client applications for both Android and iOS platforms. It also provides an extensible framework that you can use to develop your own build rules. Package: bazel Architecture: ppc64el Version: 0.4.3-3ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 120551 Depends: libc6 (>= 2.17), libgcc1 (>= 1:3.0), libstdc++6 (>= 5.2) Filename: dists/xenial/main/binary-ppc64el/bazel_0.4.3-3ibm1_ppc64el.deb Size: 119520900 MD5sum: 5229ce3c8dececef82eabd0315072a51 SHA1: ee9b313a86d4f74fe749e6e6fb3471cceb365482 SHA256: 4e203261723d613281246f1f32151c53fe81b68dfe24e8592a503c60b745b3ca SHA512: 9339ef3e43940596a6d1ed57ec44fdb8c1027bbb713dde41ef6d82c780c596f7c3578b30730ae3fc38803a092d0f718c52439d8137fb852796448a1f19db52d9 Homepage: http://www.bazel.io/ Description: Correct, reproducible, fast builds for everyone Bazel is Google's own build tool, now publicly available in Beta. Bazel has built-in support for building both client and server software, including client applications for both Android and iOS platforms. It also provides an extensible framework that you can use to develop your own build rules. Package: torch Architecture: ppc64el Version: 7-3ibm3 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 864220 Depends: gnuplot-x11, gnuplot, graphicsmagick, imagemagick, libgraphicsmagick1-dev, ipython, libfftw3-single3, python-zmq, sox, cuda-cublas-8-0, cuda-cudart-8-0, cuda-cusparse-8-0, libc6 (>= 2.22), libgcc1 (>= 1:3.3), libgomp1 (>= 4.9), libjpeg8 (>= 8c), liblmdb0 (>= 0.9.7), libopenblas, libpng12-0 (>= 1.2.13-4), libqt4-svg (>= 4:4.5.3), libqtcore4 (>= 4:4.8.4), libqtgui4 (>= 4:4.6.1), libreadline6 (>= 6.0), libssl1.0.0 (>= 1.0.0), libstdc++6 (>= 5.2), libx11-6, libzmq5 (>= 4.1.2) Recommends: ipython-notebook Filename: dists/xenial/main/binary-ppc64el/torch_7-3ibm3_ppc64el.deb Size: 40598554 MD5sum: a79f2cfd8dbaeb039dfdcd9028f83a22 SHA1: 96afa3ccce8ece6c595077f6b1a702c8fb518423 SHA256: 23d329506676b2386191935eee1c8a4e7c09fcf029530f0d70aaf3317155634b SHA512: 8601811c49d442efafee2cf956681d2359199731bef83cf35f0502d8f8cd5f4762e0361b9debaf7ae560d2b764ddc68e7c6a9b6f84d768acd3ea697c7070d1b8 Homepage: http://torch.ch Description: A scientific computing framework for LuaJIT Torch is a scientific computing framework with wide support for machine learning algorithms that puts GPUs first. It is easy to use and efficient, thanks to an easy and fast scripting language, LuaJIT, and an underlying C/CUDA implementation. Package: caffe-nv Architecture: ppc64el Version: 0.15.14-3ibm1 Priority: optional Section: science Maintainer: IBM MLDL Installed-Size: 93847 Depends: build-essential, cmake, curl, g++, gcc, git-core, protobuf-compiler, python (>= 2.7), python-skimage, python-protobuf, libnccl1, cuda-cublas-8-0, cuda-cudart-8-0, cuda-curand-8-0, libboost-filesystem1.58.0, libboost-python1.58.0, libboost-system1.58.0, libboost-thread1.58.0, libc6 (>= 2.17), libcudnn5-dev, libgcc1 (>= 1:3.0), libgflags2v5, libgoogle-glog0v5, libhdf5-10, libleveldb1v5, liblmdb0 (>= 0.9.7), libopenblas, libopencv-core2.4v5, libopencv-highgui2.4v5, libopencv-imgproc2.4v5, libprotobuf9v5, libpython2.7 (>= 2.7), libstdc++6 (>= 5.2) Conflicts: caffe Filename: dists/xenial/main/binary-ppc64el/caffe-nv_0.15.14-3ibm1_ppc64el.deb Size: 15159500 MD5sum: 4776fbb65139590a86b6f1721c9c5743 SHA1: 97226db6b497778cc22c56878b41a7f868775dd4 SHA256: 34adf718b1f95b3e9def4719a4c71745bbdf2ca0d21c7e7ad4874a84912bae72 SHA512: 11a51108a61a6022268a3a08bdbe7826127e9ed97fb4452f37cbd385bec9fc42cb663e713cc57f6c4b0dff1b053ff81e309e9c5dc7f4547919142e9028f3de94 Homepage: http://caffe.berkeleyvision.org/ Description: Fast open framework for deep learning (NVIDIA's fork) Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and by community contributors. . This package includes NVIDIA's fork of upstream Caffe.