sampling : add support for saving/loading backend sampling state #18862
+320
−47
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit updates output_reserve in llama-context.cpp to always allocate sampling buffers regardless of whether sampling is needed for the current batch. The motivation for this is to avoid reallocations and branching based on the sampling requirements of the batch.
This commit adds write/read support for backend sampling state similar to how the logits and embedding buffers are handled. The motivation for this is that it adds the backend sampling state to be saved/restored along with the rest of the llama_context state.
Member
|
Initially, I was thinking that since the samplers can now be part of the context state, we should also store this information. But it should include also the sampler states. And it gets very complicated. But now I am wondering if we should instead remove the output ids, the logits and the embeddings from the state and only store the model info and the memory. I can't think of a meaningful use case for storing this information. And even if its needed, one can simply run the last token through |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This commit adds write/read support for backend sampling state similar
to how the logits and embedding buffers are handled.
The motivation for this is that it adds the backend sampling state to
be saved/restored along with the rest of the llama_context state.
This commit build upon #18811 which is included as the first commit in this PR. I'll rebase and remove it once it has been reviewed and merged.