kubeflow/pytorch-operator

Issues

• unable to build image for ppc64le

  #365 opened 3 years ago by gajanankulkarni-18
  0

• PytorchJob DDP training will stop if I delete a worker pod

  #364 opened 3 years ago by Shuai-Xie
  2

• What is the difference between master and worker?

  #342 opened 3 years ago by SeibertronSS
  6

• Multi-gpu in a single pod

  #362 opened 3 years ago by wallarug
  2

• run https://github.com/kubeflow/pytorch-operator/blob/master/sdk/python/test/test_e2e.py failed

  #363 opened 3 years ago by sxl1993
  1

• How to use DDP in pytorch operator？

  #350 opened 3 years ago by SeibertronSS
  3

• whether multi-gpu-per-pod setup be supported in PytorchJob

  #331 opened 4 years ago by tingweiwu
  1

• service label mismatches selector, which result in inconsistency

  #360 opened 3 years ago by konnase
  3

• The training hangs after reloading one of master/worker pods

  #359 opened 3 years ago by dmitsf
  5

• Is python sdk still being maintained?

  #317 opened 4 years ago by ca-scribner
  7

• container "pytorch" is waiting to start: PodInitializing

  #348 opened 3 years ago by gogogwwb
  20

• Can not use volcano for Gang Scheduling

  #358 opened 3 years ago by bug-developer021
  0

• Can I freeze pytorchjob training pods and migrate them to other nodes?

  #356 opened 3 years ago by Shuai-Xie
  9

• Pytorch version may have an effect on the training reproduction

  #355 opened 3 years ago by Shuai-Xie
  4

• Different DDP training results of PytorchJob and Bare Metal

  #354 opened 3 years ago by Shuai-Xie
  6

• Can I use hostNetwork to run PytorchJob like on bare metal

  #353 opened 3 years ago by Shuai-Xie
  3

• Can PytorchJob skip or cancel the init cantainer?

  #352 opened 3 years ago by SeibertronSS
  2

• Mnist dataset server is down

  #325 opened 4 years ago by Jeffwan
  5

• volcano change the PodGroup CRD APIGroup to volcano.sh

  #351 opened 3 years ago by qiankunli
  1

• why worker need initContainer in pytorch-operator?

  #349 opened 3 years ago by zqz-net
  2

• [feat] Support PyTorch 1.9

  #346 opened 3 years ago by gaocegege
  3

• Upgrade to v1 CRDs

  #347 opened 3 years ago by mcristina422
  1

• kubeflow pipelines sdk, distributed multi-node training with autoscaling

  #312 opened 3 years ago by rami3e
  4

• PytorchJob replicas has different node affinity behaviors compared with Deployment

  #344 opened 3 years ago by Shuai-Xie
  4

• fell confused about world_size

  #340 opened 3 years ago by ldd91
  0

• `init-pytorch` init container image configurable

  #339 opened 4 years ago by apatil4
  4

• PyTorch Lightning Example.

  #334 opened 4 years ago by tchaton
  0

• Unlable to spawn PyTorchJob due to image alpine dependency of pytorch-operator

  #319 opened 4 years ago by asahalyft
  4

• './pytorch_job_sendrecv.yaml' missing in pytorch-operator/examples/smoke-dist

  #291 opened 4 years ago by Lyken17
  6

• can I use PyTorchJobClient inside a pod of the cluster?

  #330 opened 4 years ago by omlomloml
  1

• Worker template should be configurable.

  #335 opened 4 years ago by MartinForReal
  1

• 'host not found' error occurs during PyTorch distributed learning

  #333 opened 4 years ago by JGoo1
  1

• NCCL "Connection Refused" for Worker Pods

  #332 opened 4 years ago by twolffpiggott
  1

• worker get connection timed out error in user namespace with sidecar.istio.io/inject=false

  #329 opened 4 years ago by tingweiwu
  1

• is there a simpler way to install pytorch-operator

  #328 opened 4 years ago by tingweiwu
  2

• Please create v1.2-branch

  #314 opened 4 years ago by SatwikBhandiwad
  3

• pytorch-operator: Consolidate manifests

  #322 opened 4 years ago by yanniszark
  1

• PyTorch Operator: Move manifests development upstream

  #318 opened 4 years ago by yanniszark
  0

• Operator has invalid memory address error on specific pytorchjob spec

  #321 opened 4 years ago by ca-scribner
  1

• pytorch-operator pod CheckCRDExist failed

  #294 opened 4 years ago by myonlyzzy
  3

• dist.init_process_group stuck

  #313 opened 4 years ago by ravenj73
  9

• Does pytorch-opterator just simplified the use of nn.parallel.DistributedDataParallel on multi nodes of multi gpu？

  #311 opened 4 years ago by lwj1980s
  2

• can I use gpus on specific node to train

  #310 opened 4 years ago by lwj1980s
  5

• how can I run a pytorch job with all my Gpu resources

  #304 opened 4 years ago by lwj1980s
  4

• Make manifest test friendly

  #302 opened 4 years ago by Jeffwan
  2

• Do not trigger presubmit jobs for simple changes

  #307 opened 4 years ago by Jeffwan
  1

• Support Torch Elastic in pytorch operator

  #296 opened 4 years ago by Jeffwan
  2

• Activate Travis in PR check

  #299 opened 4 years ago by andreyvelich
  2

• [bug] Unit test is broken

  #292 opened 4 years ago by gaocegege
  4

• how to create a local non-distributed training

  #287 opened 5 years ago by houz42
  7