Dear Professor, I'm quite interested in your work and am currently replicating your code.
Theoretically, when we compress a large model using the KV-CACHE context cache algorithm, shouldn't we save the compressed model at the end? I don't see any code in your code for saving the compressed file.
So, I'd like to discuss this with you and ask for your advice.
Dear Professor, I'm quite interested in your work and am currently replicating your code.
Theoretically, when we compress a large model using the KV-CACHE context cache algorithm, shouldn't we save the compressed model at the end? I don't see any code in your code for saving the compressed file.
So, I'd like to discuss this with you and ask for your advice.