Feature: Added api for getting/setting the complete state: rng, logits, embedding and kv_cache #1105
+133
−1
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
including rng, logits, embedding and kv_cache
xaedes
force-pushed
the
state_persistence
branch
from
cab1fe0 to
1c51e1f
Compare
ggerganov
reviewed
Apr 22, 2023
|
|
||
| // Returns the size of the state | ||
| size_t llama_get_state_size(struct llama_context * ctx) { | ||
| const size_t s_bool = sizeof(int32_t); |
There was a problem hiding this comment.
s_bool is unused - is this expected?
There was a problem hiding this comment.
Oh, I missed that during cleanup of unused stuff.
At one time during implementation I was saving the bool flags from llama_context, but removed it because it didnt make much sense.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
I have implemented functions for getting and setting the rest of the model state.
It includes: random number generator state, logits, embedding and kv_cache.
It was necessary to store the logits so that we can
eval tokens,save state,restart program,load stateand thensample.With just restoring kv_cache the sampling did not have access to the required logits and indeed segfaulted on the initially empty logits vector.
The logits vector initial capacity was reserved with a wrong value. This resulted in changing capacity after the first evaluation in which the logits vector is actually resized. I fixed this bug because it propagated to the state size, resulting in unnecessarily changes.
The random number generator state is also included to ensure consistent sampling results.
Since the internal state of the rng is more than just the seed, it is serialized using the standard C++ api for this purpose by streaming into a stringbuffer. For simplicity I did not add further logic to parse and compress the serialized rng state.
For completeness I also stored the embedding vector.
Because the whole state is not in one contiguous memory buffer I decided on an output pointer parameter to get the state data.
The user is responsible to allocate the memory where the state is written to. To support this the required number of bytes can be requested.