This repository has been archived by the owner on May 28, 2024. It is now read-only.
Merge from master #25
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
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Max Pumperla <[email protected]>
…th `anyscale service rollout`
Signed-off-by: Antoni Baum <[email protected]>
Aviary currently combines all `LLMDeployments` and the `Router` deployment into a single Serve app. When a request arrives, the `Router` determines which `LLMDeployment` should process it and it sends the request to that deployment over a Serve handle. However, due to a limitation in Ray and Ray Serve, deployments called from other deployments using a Serve Handle cannot stream their responses. This means the current single-app architecture cannot support streaming. This change enables streaming by splitting the `Router` and each `LLMDeployment` into its own Serve app. To work around the limitation, whenever the `Router` app receives an HTTP request, it creates a new HTTP request over `localhost` to a streaming `LLMDeployment`, receives the response stream, and sends a `StreamingResponse` containing that stream back to the original client. This workaround is meant to be temporary and should be replaced with a longer-term solution once the Ray Serve limitation is fixed. --------- Signed-off-by: Max Pumperla <[email protected]> Signed-off-by: Shreyas Krishnaswamy <[email protected]> Signed-off-by: Antoni Baum <[email protected]> Co-authored-by: Max Pumperla <[email protected]> Co-authored-by: Kyle Huang <[email protected]> Co-authored-by: M Waleed Kadous <[email protected]> Co-authored-by: sijie <[email protected]> Co-authored-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
This change adds `locustfiles` that can load test Aviary. --------- Signed-off-by: Antoni Baum <[email protected]> Signed-off-by: Shreyas Krishnaswamy <[email protected]> Co-authored-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Signed-off-by: Antoni Baum <[email protected]>
Yard1
approved these changes
Jul 20, 2023
avnishn
pushed a commit
that referenced
this pull request
Aug 4, 2023
Signed-off-by: Antoni Baum <[email protected]>
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.