Uploaded image for project: 'Apache MXNet (Retired)'
  1. Apache MXNet (Retired)
  2. MXNET-1234

Activation gradient (backward pass) has bugs on shape inference

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Done
    • Major
    • Resolution: Fixed
    • Apache MXNet Backend
    • None

    Description

      Shape inference is not working well for the Activation backward pass and triggering assertions since it has different number of inputs depending on the type of activation such as relu, softsign etc. This logic is not correctly handled and the C++ fail on some build configurations (CPU/MKL/GPU+MKL/GPU etc).

      Attachments

        Activity

          People

            Unassigned Unassigned
            larroy Pedro Larroy
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Time Tracking

                Estimated:
                Original Estimate - Not Specified
                Not Specified
                Remaining:
                Remaining Estimate - 0h
                0h
                Logged:
                Time Spent - 6h 40m
                6h 40m