Join GitHub today
GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.Sign up
This is very related to #23132 (but not the same).
This is problematic when working on relatively large data, as we have to resort to working with raw pointers and unmanaged memory allocation to work with contiguous arrays of 2G+ elements.
To give a concrete example, from .NET I can't simply give TensorFlow 800_000 64x64 1-channel (channel==byte) images as a contiguous array.