0.2.4

Highlights

• Compatibility with vLLM v0.5.4
• Bump minimum vLLM requirement to v0.5.4

What's Changed

• remove dead code (vllm<=0.5.0.post1) by @dtrifiro in #65
• build(deps): bump ruff from 0.5.4 to 0.5.5 by @dependabot in #72
• pyproject: fix broken URLs by @dtrifiro in #76
• gha: add missing build dependencies by @dtrifiro in #68
• noxfile: make overridden vllm version install verbose by @dtrifiro in #78
• updates for vLLM==0.5.4 by @dtrifiro in #82
• fix merge_async_iterators usage for vLLM>0.5.4 by @dtrifiro in #86
• build(deps): bump hf-transfer from 0.1.6 to 0.1.8 by @dependabot in #73
• build(deps): bump flash-attn from 2.6.1 to 2.6.3 by @dependabot in #71
• pre-commit: bump deps by @dtrifiro in #67
• Extract request's trace context in GenerateStream() by @ronensc in #64
• Make ADD_SPECIAL_TOKENS true by default by @maxdebayser in #66
• build(deps): bump pytest from 8.2.2 to 8.3.2 by @dependabot in #70
• Fix stop_reason for secondary eos tokens by @njhill in #75
• ♻️ refactor LoRARequest based on upstream by @prashantgupta24 in #74
• build(deps): bump mypy from 1.10.1 to 1.11.1 by @dependabot in #77
• Fix length penalty logit processor getting NaN by @wallashss in #85
• Re-enable seed for pipeline parallel by @njhill in #69
• 🥅 Kill servers on engine death by @joerunde in #63

New Contributors

• @ronensc made their first contribution in #64
• @maxdebayser made their first contribution in #66
• @wallashss made their first contribution in #85

Full Changelog: 0.2.3...0.2.4

0.2.3

What's Changed

• build(deps): bump accelerate from 0.31.0 to 0.32.1 by @dependabot in #42
• build(deps): bump flash-attn from 2.5.9.post1 to 2.6.1 by @dependabot in #45
• Fix earlier LoRA tokenizer changes by @njhill in #53
• build(deps): bump ruff from 0.5.1 to 0.5.2 by @dependabot in #43
• build(deps): bump types-requests from 2.32.0.20240602 to 2.32.0.20240712 by @dependabot in #44
• Fix tokenize endpoint by @njhill in #54
• build(deps): bump ruff from 0.5.2 to 0.5.4 by @dependabot in #58
• vLLM 5.3+ support by @joerunde in #60

New Contributors

• @joerunde made their first contribution in #60

Full Changelog: 0.2.2...0.2.3

0.2.2

Full Changelog: 0.2.1...0.2.2

What's Changed

• tests: fix lora tests handling by @dtrifiro in #51
• fix missing metrics endpoint by @dtrifiro in #50

Full Changelog: 0.2.1...0.2.2

0.2.1

What's Changed

• update for 0.5.2 by @dtrifiro in #46
• deps: pin transformers by @dtrifiro in #47
• Fix use of LoRA tokenizers by @njhill in #40

Full Changelog: 0.2.0...0.2.1

0.2.0

What's Changed

• improve custom vllm version testing by @dtrifiro in #33
• build(deps): bump mypy from 1.10.0 to 1.10.1 by @dependabot in #27
• build(deps): bump ruff from 0.4.9 to 0.5.1 by @dependabot in #34
• add support for prompt adapters by @prashantgupta24 in #21
• Add mapping for TGIS speculator args by @njhill in #39

New Contributors

• @njhill made their first contribution in #39

Full Changelog: 0.1.3...0.2.0

0.1.3

What's Changed

• improve tasks handling by @dtrifiro in #28
• fix convert output by @dtrifiro in #29
• gha: test multiple versions by @dtrifiro in #30
• fix main module by @dtrifiro in #31
• gha: fix coverage report format by @dtrifiro in #32

Full Changelog: 0.1.2...0.1.3

0.1.2

What's Changed

• Apply inferface change to duplicated code in the tgis layer by @prashantgupta24 in #16
• gha: add merge_group trigger by @dtrifiro in #18
• add tokenization with truncation, offset support (IBM #47) by @prashantgupta24 in #19
• Respect trace headers in grpc server (IBM #49) by @prashantgupta24 in #20
• build(deps): bump flash-attn from 2.5.3 to 2.5.9.post1 by @dependabot in #1
• build(deps): bump grpcio from 1.62.1 to 1.64.1 by @dependabot in #2
• build(deps): bump pytest from 8.2.0 to 8.2.2 by @dependabot in #3
• build(deps): bump accelerate from 0.28.0 to 0.31.0 by @dependabot in #17
• build(deps): bump grpcio-reflection from 1.62.1 to 1.64.1 by @dependabot in #4

Full Changelog: 0.1.1...0.1.2

0.1.1

What's Changed

• fix crashes by @dtrifiro in #15

Full Changelog: 0.1.0...0.1.1

0.1.0

What's Changed

• adapt for odh by @dtrifiro in #6
• build(deps): bump ruff from 0.4.3 to 0.4.9 by @dependabot in #8
• ✨ add ability to run REST with grpc server by @prashantgupta24 in #9
• concurrent http and grpc by @dtrifiro in #12
• integrate IBM vLLM changes by @dtrifiro in #14

New Contributors

• @dtrifiro made their first contribution in #6
• @dependabot made their first contribution in #8
• @prashantgupta24 made their first contribution in #9

Full Changelog: https://github.com/opendatahub-io/vllm-tgis-adapter/commits/0.1.0