This is a first step. It's ideal for saving across runs, running on a cluster and investigating later, etc.
It's still worthwhile, though, to have a way of saving data that allows transfer to Matlab, Python, etc. The OCaml npy lib seems like a good bet. I would rather not use hdf5 until someone fixes the high-level convenience functions. See #6
This is a first step. It's ideal for saving across runs, running on a cluster and investigating later, etc.
It's still worthwhile, though, to have a way of saving data that allows transfer to Matlab, Python, etc. The OCaml npy lib seems like a good bet. I would rather not use hdf5 until someone fixes the high-level convenience functions. See #6