Required prerequisites
Motivation
Hi, TorchOpt team!
I am writing to inquire if there is a function that converts torch.optim state checkpoint 100% to a torchopt object, that can replay completely the same resulting behavior as how the former behaves.
I have searched for function declarations like torchopt.from_torch etc, but seemed not to discover any.
Could you please point me to the function, inform me that it does not exist anywhere, or perhaps suggest a viable near-identical ways to implement this as close and convenient as possible? I'd be greatly indebted!
Solution
No response
Alternatives
No response
Additional context
No response
Required prerequisites
Motivation
Hi, TorchOpt team!
I am writing to inquire if there is a function that converts torch.optim state checkpoint 100% to a torchopt object, that can replay completely the same resulting behavior as how the former behaves.
I have searched for function declarations like torchopt.from_torch etc, but seemed not to discover any.
Could you please point me to the function, inform me that it does not exist anywhere, or perhaps suggest a viable near-identical ways to implement this as close and convenient as possible? I'd be greatly indebted!
Solution
No response
Alternatives
No response
Additional context
No response