feat: add prefix cache for KV cache reuse during inference#928
Open
enjoysport2022 wants to merge 1 commit intoPhysical-Intelligence:mainfrom
Open
feat: add prefix cache for KV cache reuse during inference#928enjoysport2022 wants to merge 1 commit intoPhysical-Intelligence:mainfrom
enjoysport2022 wants to merge 1 commit intoPhysical-Intelligence:mainfrom
Conversation
enjoysport2022
requested review from
Michael-Equi,
jimmyt857 and
kvablack
as code owners
Add caching infrastructure to avoid redundant computation of prefix embeddings (images + language) when observation hasn't changed between consecutive inference calls. This can provide significant speedup in continuous robot control scenarios. Changes: - Add prefix_cache.py with PrefixCacheKey and PrefixCache classes - Add sample_actions_with_cache() method to Pi0 model - Add enable_prefix_cache parameter to Policy and create_trained_policy - Add clear_cache() and reset() methods to Policy Usage: policy = create_trained_policy(config, checkpoint, enable_prefix_cache=True) result = policy.infer(obs) # Auto-caches prefix policy.reset() # Clear cache for new episode
enjoysport2022
force-pushed
the
feature/prefix-cache-optimization
branch
from
b4411ea to
b64ff9d
Compare
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Add caching infrastructure to avoid redundant computation of prefix embeddings (images + language) when observation hasn't changed between consecutive inference calls. This can provide significant speedup in continuous robot control scenarios.
Changes:
Usage:
policy = create_trained_policy(config, checkpoint, enable_prefix_cache=True) result = policy.infer(obs) # Auto-caches prefix policy.reset() # Clear cache for new episode