List view

• Pip package

  PyVertical should be more than just a playground/demonstration of vertical federated learning: it should be a pip-installable package which provides (a) utility functions for splitting datasets, for the purposes of VFL experiments (b) utility functions for creating Split neural networks (c) code for interfacing with syft/grid

  No due date

  •2/7 issues closed
  28% complete5 open2 closed

• Syft >= 0.5

  Tasks which will not make `syft 0.4.0`. - Multi-party (n > 2) vertical federated learning - Asymmetric vfl

  No due date

  •0/2 issues closed
  0% complete= 0_5"">2 open= 0_5"">0 closed

• Syft 0.4

  `PyVertical` is to be integrated with `syft 0.4.0`. Some key functionality is being developed in `0.4.0` independently of `PyVertical`: - split neural networks - PSI over duet - federated learning plans `PyVertical` should offer useful examples of vertical federated learning in syft and implement: - vertically partitioned datasets and dataloaders - a number of demos

  Overdue by 4 year(s)

  •

  Due by February 1, 2021

  •5/6 issues closed
  83% complete1 open5 closed

• Extended example

  After the MVP, we want to extend PyVertical to work on a more complex dataset. This milestone represents a more "real-world" application of PyVertical

  No due date

  •3/5 issues closed
  60% complete2 open3 closed