KV Cache Improved Flexibility #4668
Merged
Conversation
This file contains 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
cmikeh2
requested review from
RezaYazdaniAminabadi,
jeffra,
mrwyattii,
awan-10,
arashb and
tjruwase
as code owners
awan-10
approved these changes
Nov 14, 2023
jeffra
reviewed
Nov 16, 2023
| @@ -124,7 +116,9 @@ def flush_sequence(self, uid: int) -> None: | |||
| return | |||
|
|
|||
| seq = self._seqs[uid] | |||
| self._kv_cache.free(seq.all_block_ids) | |||
| for i in range(self.n_kv_caches): | |||
There was a problem hiding this comment.
self.n_kv_caches doesn't exist :(
jeffra
added a commit
that referenced
this pull request
Nov 16, 2023
mauryaavinash95
pushed a commit
to mauryaavinash95/DeepSpeed
that referenced
this pull request
Feb 17, 2024
This KV-cache adds the foundation for appropriately supporting two key KV-cache improvements: 1. Delineation between local/dense KV caches/models at the cache level in addition to the attention module level. 2. Support for multiple types of disjoint KV caches (such as alternating local + dense attention GPT-Neo). Follow up item: Determine appropriate statistics for weighting local + dense KV block ratios when both are present. --------- Co-authored-by: Olatunji Ruwase <[email protected]>
mauryaavinash95
pushed a commit
to mauryaavinash95/DeepSpeed
that referenced
this pull request
Feb 17, 2024
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Outdated suggestions cannot be applied.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.
This KV-cache adds the foundation for appropriately supporting two key KV-cache improvements:
Follow up item: Determine appropriate statistics for weighting local + dense KV block ratios when both are present.