Skip to content

Prefetch inefficiency

The way we do prefetch is quite inefficient, basically because of the GIL.

If someone is interested, let's observe this thread here (or even collaborate with it)

https://github.com/tensorflow/tensorflow/issues/7951

@pkorshunov @heusch @amohammadi

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information