Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / chainer/chainermn issues and pull requests

#299 - Installation should do nothing but omit a warning.

Issue - State: open - Opened by keisukefukuda over 5 years ago

#298 - Revert ReadTheDocs latest site

Pull Request - State: closed - Opened by kmaehashi over 5 years ago

#297 - Tombstone: moved onto chainer/chainer

Pull Request - State: closed - Opened by kuenishi about 6 years ago

#296 - Bump version and not allow Chainer 5.x

Pull Request - State: closed - Opened by kuenishi over 6 years ago
Labels: enhancement

#295 - Fix regression bug on cupy arrays as well as add slow annotations

Pull Request - State: closed - Opened by kuenishi over 6 years ago
Labels: bug

#294 - Modify image of parallel convolution

Pull Request - State: closed - Opened by levelfour over 6 years ago
Labels: document

#293 - Fix errors on 0-d array input to Communicator APIs

Pull Request - State: closed - Opened by kuenishi over 6 years ago
Labels: bug

#292 - Override Optimizer.setup() method at multi-node optmizers

Pull Request - State: closed - Opened by kuenishi over 6 years ago
Labels: bug

#291 - Non-Blocking Methodology on ChainerMN

Issue - State: closed - Opened by arthuryuan1987 over 6 years ago - 3 comments

#290 - Workaround forkserver

Pull Request - State: closed - Opened by kuenishi over 6 years ago
Labels: document

#289 - add mnbn with nccl

Pull Request - State: closed - Opened by shu65 over 6 years ago
Labels: feature

#288 - bugfix bcast

Pull Request - State: closed - Opened by shu65 over 6 years ago
Labels: bug

#287 - CUDA streams usage

Issue - State: closed - Opened by mshiryaev over 6 years ago - 6 comments

#286 - Update Chainer version to 4.4.0 in .travis.yml

Pull Request - State: closed - Opened by keisukefukuda over 6 years ago - 1 comment
Labels: test

#285 - NCCL_ERROR_SYSTEM_ERROR: unhandled system error

Issue - State: closed - Opened by Fhrozen over 6 years ago - 3 comments

#284 - added OMP_NUM_THREADS=1

Pull Request - State: closed - Opened by keisukefukuda over 6 years ago
Labels: test

#283 - When `in_size=None` is used in `Liner` and it is not used, an error occurs

Issue - State: closed - Opened by shu65 over 6 years ago
Labels: bug

#282 - Reduce CUDA kernel launch in BN (updated)

Pull Request - State: closed - Opened by keisukefukuda over 6 years ago - 2 comments
Labels: enhancement

#281 - Manual cherry-pick of fda23e482d37321

Pull Request - State: closed - Opened by kuenishi over 6 years ago
Labels: document

#280 - Remove redundant "would" in installation section

Pull Request - State: closed - Opened by nobu-k over 6 years ago - 1 comment
Labels: document

#279 - Travis update

Pull Request - State: closed - Opened by keisukefukuda over 6 years ago
Labels: test

#278 - Forcing forkserver spawn earlier

Issue - State: closed - Opened by iwiwi over 6 years ago - 2 comments
Labels: document

#277 - FP16 support

Issue - State: open - Opened by kuenishi over 6 years ago - 1 comment
Labels: feature

#276 - Reduce CUDA kernel launch in BN

Pull Request - State: closed - Opened by okuta over 6 years ago - 1 comment

#275 - optimizer.setup() created by create_multi_node_optimizer returns an original optimizer

Issue - State: closed - Opened by rezoo over 6 years ago - 2 comments
Labels: bug

#274 - Add `force_equal_length` flag to `scatter_dataset` method

Issue - State: open - Opened by iwiwi over 6 years ago
Labels: feature

#273 - CommunicatorBase.{scatter, allgather} is missing in the document

Issue - State: open - Opened by iwiwi over 6 years ago
Labels: document

#272 - Add parallel convolution example

Pull Request - State: closed - Opened by levelfour over 6 years ago
Labels: example

#271 - Bugfix bcast for FP16

Pull Request - State: closed - Opened by shu65 over 6 years ago
Labels: bug

#270 - Improve performance of fetching device memory

Pull Request - State: closed - Opened by levelfour over 6 years ago - 4 comments
Labels: enhancement

#269 - Manual selection for gpus in distributed training

Issue - State: closed - Opened by 1292765944 over 6 years ago - 5 comments

#268 - update tested env

Pull Request - State: closed - Opened by shu65 over 6 years ago

#266 - Bump version and year

Pull Request - State: closed - Opened by kuenishi over 6 years ago

#265 - Enable MultiNodeIterator/MpiCommunicatorBase to handle general data type

Pull Request - State: closed - Opened by levelfour over 6 years ago - 2 comments

#264 - Add note on checkpoints and fix broken autofunction

Pull Request - State: closed - Opened by kuenishi over 6 years ago
Labels: document

#263 - Expose intra- and inter- rank and size

Pull Request - State: closed - Opened by kuenishi over 6 years ago
Labels: document, feature

#262 - Update the description about using FP16

Pull Request - State: closed - Opened by kuenishi over 6 years ago
Labels: document

#261 - Update the description about using FP16

Pull Request - State: closed - Opened by shu65 over 6 years ago - 2 comments

#260 - Modify receivers in communicator to use cupy array

Pull Request - State: closed - Opened by levelfour over 6 years ago - 2 comments

#259 - Refactor MultiNodeIterator classes

Pull Request - State: closed - Opened by shu65 over 6 years ago

#258 - Provide functions for allreduce

Issue - State: open - Opened by kuenishi over 6 years ago
Labels: feature

#257 - Remove unused nccl comm and mpi comm

Pull Request - State: closed - Opened by shu65 over 6 years ago
Labels: enhancement

#256 - Remove unused nccl comm and mpi comm

Pull Request - State: closed - Opened by shu65 over 6 years ago

#255 - Expose `intra_size`, `inter_rank` and `inter_size` of communicators at readthedocs

Issue - State: closed - Opened by iwiwi over 6 years ago
Labels: document

#254 - would you please share hype parameters of GPUs=4 for resnet50 training with us ?

Issue - State: closed - Opened by mingxiaoh over 6 years ago - 23 comments

#253 - Fix errors in initialization of NCCL

Pull Request - State: closed - Opened by shu65 over 6 years ago - 1 comment

#252 - Handle list of dicts in MultiNodeIterator

Issue - State: closed - Opened by kuenishi over 6 years ago - 1 comment

#251 - Refactor optimizers_tests

Pull Request - State: closed - Opened by shu65 over 6 years ago

#250 - Add a global exception handler to call MPI_Abort

Pull Request - State: closed - Opened by keisukefukuda over 6 years ago - 3 comments

#249 - Added an FAQ entry about MPI hang issue.

Pull Request - State: closed - Opened by keisukefukuda over 6 years ago - 3 comments
Labels: document

#248 - Fix MultiNodeIterator for paired datasets

Pull Request - State: closed - Opened by levelfour over 6 years ago - 3 comments

#247 - Dummy PR

Pull Request - State: closed - Opened by shu65 over 6 years ago

#246 - Fix tests of MultiNodeIterator

Pull Request - State: closed - Opened by shu65 over 6 years ago

#245 - [WIP] Revert multi-node iterator

Pull Request - State: closed - Opened by kuenishi over 6 years ago - 1 comment

#244 - [wip] Eliminate mpi4py's ssend to fix tests

Pull Request - State: closed - Opened by kuenishi over 6 years ago

#243 - Fix p2p-communication test

Pull Request - State: closed - Opened by kuenishi over 6 years ago

#242 - Re: Add collective communications

Pull Request - State: closed - Opened by keisukefukuda almost 7 years ago

#241 - Asynchronous Allreduce

Issue - State: closed - Opened by arthuryuan1987 almost 7 years ago - 2 comments

#240 - Remove unused nccl comm and mpi comm

Pull Request - State: closed - Opened by shu65 almost 7 years ago - 1 comment

#239 - Fix PR235

Pull Request - State: closed - Opened by shu65 almost 7 years ago - 1 comment

#238 - Update supported Chainer versions

Pull Request - State: closed - Opened by kuenishi almost 7 years ago
Labels: enhancement

#237 - Add allreduce method to communicator interface with implementation

Pull Request - State: closed - Opened by kuenishi almost 7 years ago
Labels: feature

#236 - mpirun doesn't exit when exception is thrown in some process

Issue - State: closed - Opened by andremoeller almost 7 years ago - 7 comments

#235 - Expose CommunicatorBase as communicator interface with docs

Pull Request - State: closed - Opened by kuenishi almost 7 years ago - 2 comments
Labels: enhancement, document

#234 - Adding allreduce for ndarray

Issue - State: closed - Opened by Hakuyume almost 7 years ago - 10 comments
Labels: feature

#233 - Fix a bug of NStepRNN

Pull Request - State: closed - Opened by shu65 almost 7 years ago

#232 - Clean up Communicator interface with changes

Pull Request - State: closed - Opened by kuenishi almost 7 years ago - 1 comment
Labels: enhancement

#231 - Replace get_device

Pull Request - State: closed - Opened by shu65 almost 7 years ago - 4 comments
Labels: enhancement

#230 - [WIP] Refactor Communicators

Pull Request - State: closed - Opened by shu65 almost 7 years ago

#229 - Fix bcast

Pull Request - State: closed - Opened by levelfour almost 7 years ago - 1 comment

#228 - A dummy PR

Pull Request - State: closed - Opened by keisukefukuda almost 7 years ago

#227 - PR for test scripts debug

Pull Request - State: closed - Opened by shu65 almost 7 years ago

#226 - Add collective communications

Pull Request - State: closed - Opened by levelfour almost 7 years ago - 2 comments

#225 - Checkpointer doesn't resume current learning rate

Issue - State: closed - Opened by Guriido almost 7 years ago - 8 comments

#224 - Don't inicialize global NCCL comm when

Issue - State: closed - Opened by undertherain almost 7 years ago - 2 comments
Labels: bug

#223 - Update chainer version 4.0.0rc1 / 3.5

Pull Request - State: closed - Opened by keisukefukuda almost 7 years ago
Labels: enhancement

#222 - Fix MultiNodeNStepRNN to use Chainer n_cells

Pull Request - State: closed - Opened by levelfour almost 7 years ago
Labels: bug

#221 - ChainerMN hangs with Open MPI 3

Issue - State: closed - Opened by keisukefukuda almost 7 years ago - 1 comment

#219 - Test the combination of MutliNodeIterator and MultiprocessIterator

Pull Request - State: closed - Opened by levelfour almost 7 years ago - 1 comment

#217 - Multi-GPU training hangs

Issue - State: closed - Opened by andremoeller almost 7 years ago - 14 comments

#215 - [WIP] Warn mp start method

Pull Request - State: closed - Opened by keisukefukuda almost 7 years ago - 2 comments

#214 - Fix send to avoid deadlock without inputs does not reqires grad

Pull Request - State: closed - Opened by levelfour almost 7 years ago - 1 comment
Labels: bug

#213 - Check contiguousness of outgoing arrays

Pull Request - State: closed - Opened by levelfour almost 7 years ago - 1 comment
Labels: bug

#207 - [WIP] fix deadlock in unit tests

Pull Request - State: closed - Opened by keisukefukuda almost 7 years ago

#204 - Cannot use other start method for multiprocessing

Issue - State: open - Opened by Guriido almost 7 years ago - 11 comments

#203 - Port Chainer#4191 or use Chainer's BN implementation

Issue - State: open - Opened by kuenishi almost 7 years ago - 2 comments
Labels: bug, question

#194 - Add explanation of methods of communicator to document

Issue - State: closed - Opened by iwiwi almost 7 years ago - 1 comment

#189 - Make test output colorful

Pull Request - State: closed - Opened by keisukefukuda about 7 years ago
Labels: test

#188 - [WIP] Supress redundant test check status reports caused by MPI

Pull Request - State: closed - Opened by keisukefukuda about 7 years ago - 2 comments
Labels: test

#187 - Add FP16 and FP64 Supports to PureNcclComunicator

Pull Request - State: closed - Opened by shu65 about 7 years ago
Labels: feature

#186 - MultiNodeIterator

Pull Request - State: closed - Opened by levelfour about 7 years ago
Labels: feature

#185 - [WIP] Add Synchronized Iterator

Pull Request - State: closed - Opened by levelfour about 7 years ago - 5 comments
Labels: feature

#161 - [WIP] Try CicleCI

Pull Request - State: closed - Opened by kuenishi about 7 years ago - 1 comment

#150 - [pending] Hotfix flake8

Pull Request - State: closed - Opened by kuenishi about 7 years ago - 2 comments

#145 - ChainerMN's ImageNet example is slower than Chainer's data parallel

Issue - State: open - Opened by LWisteria about 7 years ago - 2 comments

#87 - Mention sudo's env-var issue in the installation document

Pull Request - State: closed - Opened by keisukefukuda over 7 years ago
Labels: document

#76 - Creating trainer snapshots?

Issue - State: closed - Opened by MannyKayy over 7 years ago - 3 comments
Labels: enhancement

#71 - Add Dockerfile

Pull Request - State: closed - Opened by MannyKayy over 7 years ago - 4 comments
Labels: example

#31 - FP16 support

Issue - State: closed - Opened by iwiwi over 7 years ago - 2 comments
Labels: feature