Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

integrate IBM vLLM changes #14

Merged
merged 5 commits into from
Jun 20, 2024
Merged

integrate IBM vLLM changes #14

merged 5 commits into from
Jun 20, 2024

Conversation

dtrifiro
Copy link
Contributor

  • Add guided decoding to TGIS gRPC API
  • add tgi_request_duration histogram
  • Generic adapter support in the grpc server
  • Fix logging of stop reason for streaming requests
  • replace with _get_logits_processor

@dtrifiro dtrifiro force-pushed the ibm-vllm-changes branch 2 times, most recently from af4b318 to 13b74a9 Compare June 20, 2024 13:20
@dtrifiro dtrifiro merged commit d7281cf into main Jun 20, 2024
@dtrifiro dtrifiro deleted the ibm-vllm-changes branch June 20, 2024 13:20
@dtrifiro dtrifiro mentioned this pull request Jun 20, 2024
Closed
njhill and others added 5 commits June 20, 2024 15:21
@njhill @dtrifiro
Add guided decoding to TGIS gRPC API
fe771ea 
  enum ResponseFormat {
    // Plain text, no constraints
    TEXT = 0;
    // Valid json
    JSON = 1;
  }

  message StringChoices {
    repeated string choices = 1;
  }

  // Mutually-exclusive guided decoding options
  oneof guided {
    // Output will be in the specified format
    ResponseFormat format = 3;
    // Output will follow the provided JSON schema
    string json_schema = 4;
    // Output will follow the provided regex pattern
    string regex = 5;
    // Output will be exactly one of the specified choices
    StringChoices choice = 6;
    // Output will follow the provided context free grammar
    string grammar = 7;
  }

Signed-off-by: Nick Hill <[email protected]>
@joerunde @dtrifiro
add tgi_request_duration histogram
d1d1ed7 
Adds in this one more metric from TGIS

---------

Signed-off-by: Joe Runde <[email protected]>
@joerunde @dtrifiro
Generic adapter support in the grpc server
d50b685 
---------

Signed-off-by: Joe Runde <[email protected]>
@njhill @dtrifiro
Fix logging of stop reason for streaming requests
c4d836b 
Also don't set/return seed or other (random) sampling params when in greedy mode.

Signed-off-by: Nick Hill <[email protected]>
Signed-off-by: Prashant Gupta <[email protected]>
@prashantgupta24 @dtrifiro
replace with _get_logits_processor
13b74a9 
Signed-off-by: Prashant Gupta <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@dtrifiro @njhill @joerunde @prashantgupta24