Saving runs#
The hybrid_aif.data package reads and writes episodes on disk in two formats, a folder of
arrays and HDF5. It is pure numpy. A run’s record can be written with a writer and read back with
load_dataset; the arrays it returns are what the replay facility takes. The neural training
buffers read the same formats.
Reading and writing#
Open |
|
Return the first registered format whose |
|
Base class for episode-based datasets. |
|
Streaming episode writer — append episodes, finalize on exit. |
|
Declarative format spec. |