You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello, so my issue is related to this issue, I want to do some offline things where I need to reset the simulator to past states, using the generalized backend, and it was stated that I should store the entire State.
My question is how to store this since State is a complex Python structure, the most naïve way would be to store it in a Python list, but this would be awful to sample from this Python list. For me, the best thing to do would be to convert the State structure to a list of many tensors, what do you think about this approach?
The text was updated successfully, but these errors were encountered:
Thank you! Yes, that's exactly the kind of thing I was looking for. I'm gonna use the list for the batch dimension of State.
What I meant by "awful to sample from" is the wall-clock time of the operation, I may be wrong, but I think wall-clock time is way better for sampling from tensors structures with libraries such as numpy rather than sampling from Python data structures that are agnostic to the data structure like deque or list. That's why in the end I'm gonna use the tree_leaves to discard the tree structure as you suggested, but reconstruction seems a little bit painful.
Hello, so my issue is related to this issue, I want to do some offline things where I need to reset the simulator to past states, using the generalized backend, and it was stated that I should store the entire State.
My question is how to store this since State is a complex Python structure, the most naïve way would be to store it in a Python list, but this would be awful to sample from this Python list. For me, the best thing to do would be to convert the State structure to a list of many tensors, what do you think about this approach?
The text was updated successfully, but these errors were encountered: