Add documentation and sample on KV cache eviction #1960
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
github-actions
bot
added
category: continuous batching
site/docs/concepts/optimization-techniques/kvcache-eviction-algorithm.md
Show resolved
Hide resolved
site/docs/concepts/optimization-techniques/kvcache-eviction-algorithm.md
Outdated
Show resolved
Hide resolved
site/docs/concepts/optimization-techniques/kvcache-eviction-algorithm.md
Outdated
Show resolved
Hide resolved
yatarkan
reviewed
Mar 26, 2025
| ## Conceptual Model | ||
| The KV cache for each sequence is divided into three logical areas: | ||
|
|
||
|  |
There was a problem hiding this comment.
Images should be placed at site/static/img directory and referenced here from this directory:
Suggested change
|  | |
|  |
AlexKoff88
reviewed
Mar 26, 2025
| @@ -0,0 +1,247 @@ | |||
|
|
|||
| import gc | |||
There was a problem hiding this comment.
I think this is an example that should be moved to the samples folder.
l-bat
reviewed
Mar 26, 2025
github-actions
bot
added
category: samples
AlexKoff88
reviewed
Mar 27, 2025
|
|
||
| * Start Area: Initial tokens that are never evicted | ||
| * Evictable Area: Tokens that can be evicted based on importance scores | ||
| * Recent Area: Most recent tokens that are preserved (never evicted) |
There was a problem hiding this comment.
Suggested change
| * Recent Area: Most recent tokens that are preserved (never evicted) | |
| * Recent Area: Most recent tokens that are preserved but migrate to the Evictable Area with iterations of text generation. |
|
Well-written! I would add a paragraph at the end like "Areas/Subjects for improvements". |
Done |
AlexKoff88
approved these changes
Mar 28, 2025
l-bat
approved these changes
Mar 28, 2025
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.
No description provided.