Labels

• Affects linear_operator refactor
• batch mode
• breaking change
• bug
• compatibility
• derivatives

  Questions about GP+derivative models
• discussion
• DKL

  For questions about DKL
• documentation
• duplicate
• enhancement
• extending gpytorch

  When people are trying to write custom kernels, etc.
• fantasy points
• good first issue
• help wanted
• high priority
• invalid
• major undertaking
• memory/efficiency
• multitask

  For questions about multitask models
• push upstream

  Functionality that is generally useful and should be moved into pytorch
• question
• refactor
• stability

  When models return NaNs and stuff
• unit tests
• update_on_next_release

  These are features that are compatible with stable/bleeding-edge versions of PyTorch.
• variational

  For questions about variational inference
• WIP
• wontfix