huggingface/text-generation-inference

Pull Requests Commits

feat: consolidate streaming and event creation logic and add tests for streaming generations

drbh committed 1 year ago

efb20054

feat: improve partial parsing types and add test for balancing and partial parsing

drbh committed 1 year ago

330f2e41

feat: refactor and simplify chat stream more, bump tests and support stream_options

drbh committed 1 year ago

a5ddc9db

fix: bump integrations requirements

drbh committed 1 year ago

c4cb54c2

feat: refactor chat stream to remove state machine and simplfy logic

drbh committed 1 year ago

31a536d7

fix: adjust integration tests for openai client dep

drbh committed 1 year ago

a416ddbd

fix: clippy cleanup

drbh committed 1 year ago

e1b6d5be

fix: only send function name on first stream event

drbh committed 1 year ago

538456ba

fix: bump requirements file too

drbh committed 1 year ago

68aa6b1a

fix: bump integration test deps for openai

drbh committed 1 year ago

fd611f30

fix: bump client tests for api changes

drbh committed 1 year ago

7d17d7ce

fix: bump client test expected prefill

drbh committed 1 year ago

c215c0de

fix: remove snap with incorrect naming

drbh committed 1 year ago

1529a676

fix: adjust stream, improve tests and add openai client test

drbh committed 1 year ago

40f905d0

fix: ensure wrapping curly is not included

drbh committed 1 year ago

07c20903

fix: adjust streaming tool response

drbh committed 1 year ago

dbce04e4

fix: bump openapi spec

drbh committed 1 year ago

5f030140

feat: serialize function definition with serialize_as_string

drbh committed 1 year ago

0ca7af88

fix: Functioncall is actually a bit different than the deprecated function definition type

trofleb committed 1 year ago

983b9675

fix: Allow back arguments in function definition and the corresponding test

trofleb committed 1 year ago

c1c4dfb5

feat: Make streaming for tool calling behave the same as the open ai api

trofleb committed 1 year ago

8542e2b7

fix: Adapt function call response to return a json string for arguments

trofleb committed 1 year ago

9a9a763e

Avoid running neuron integration tests twice (#3054)

dacorvo committed 1 year ago

Verified 5eec3a8b

fix: run linters and fix formatting (#3057)

drbh committed 1 year ago

Verified b0069e04

some minor fix (#3048)

sywangyi committed 1 year ago

Verified d7a24c03

You need to seek apparently. (#3049)

Narsil committed 1 year ago

Verified cea9dbc9

Add Neuron backend (#3033)

dacorvo committed 1 year ago

Verified c00add9c

Use `rotary` kernel from the Hub (#3041)

danieldk committed 1 year ago

Verified 97c5f7e6

Improve tool call message processing (#3036)

drbh committed 1 year ago

Verified 1cae3197

Update Gradio ChatInterface configuration in consuming_tgi.md (#3042)

angt committed 1 year ago

Verified 3498f608

Older