Repository navigation
What's Changed
- integrate v0.33 trtllm into master by @lxning in #2925
- [docker] Update vllm nightly wheel by @xyang16 in #2946
- Updated s5cmd script to build with specified version of go to address related CVEs by @smouaa in #2945
- Fix integration tests by @xyang16 in #2947
- [ci] Disable TestAarch64 by @xyang16 in #2948
- [ci] Fix aarch64 docker publish pipeline by @xyang16 in #2950
- Fix race condition in LoRA streaming requests by @Lokiiiiii in #2954
- Update vLLM to version 0.11.1 by @ksuma2109 in #2956
- Adding back adapter request registration by @Lokiiiiii in #2959
- Install vLLM from Extensions package wheel by @ksuma2109 in #2961
- Release notes for LMIv17 by @ksuma2109 in #2964
- Update vLLM user guide with Sample configs by @ksuma2109 in #2966
- Support for adapter-specifc pre/post processing through the adapter management APIs by @Lokiiiiii in #2951
- Fix GitHub Actions script injection vulnerabilities by @Lokiiiiii in #2967
- Add agent steering guidelines by @Lokiiiiii in #2968
- Add adapter security validation to Secure Mode plugin by @Lokiiiiii in #2970
- [ci] Use serving_version in nightly pipeline by @xyang16 in #2973
- [ci] Use serving_version in CI by @xyang16 in #2975
- Upgrade vllm to 0.12.0 and fix Multimodal integration tests by @ksuma2109 in #2978
- Lmcache benchmarks by @smouaa in #2971
- Add LMCache Benchmark tests by @ksuma2109 in #2980
- [lmi v18] Upgrade DJL and DJL-Serving to 0.36.0 by @ethnzhng in #2981
- ci: scope down GitHub Token permissions by @AdnaneKhan in #2979
- Build Docker images from java source code in release mode by @HappyAmazonian in #2984
- Add LoRA accuracy validation to adapter integration tests by @HappyAmazonian in #2985
- LCNC lmcache auto config by @smouaa in #2982
- Update LMI container to PyTorch 2.9.1, vLLM 0.14.0, and protobuf 6.30+ by @smouaa in #2988
- updated vllm toolparser imports that changed in 0.14.0 + fixed LMCache integration tests by @smouaa in #2989
- test_lora_llama3_8b_async_with_custom_code - increase output length by @smouaa in #2990
- Minor fixes for lmi release by @smouaa in #2992
- Make custom @output_formatter the final LMI response by @ksuma2109 in #2993
- Upgrade vllm to 0.15.0 with DeepSeek and GPT OSS Eagle Regression Fix by @ksuma2109 in #2995
- Update README.md by @johnpoth in #2994
- [python] Fix formatting by @xyang16 in #2997
- add release notes for v19 by @smouaa in #2998
- [docker] Upgrade to cuda 12.9 by @xyang16 in #2996
- Release notes for LMIv20 by @ksuma2109 in #2999
- updating LMIv21 container requirements to use updated wheel by @access2rohit in #3001
- fix docker build error by freezing setuptools version by @access2rohit in #3002
- Add tuned configs for LMIv21 and new vllm 0.15.1 wheel by @access2rohit in #3003
- upgrade vllm to v0.16.0 by @access2rohit in #3004
- vllm renamed AnyTokenizer to TokenizerLike. by @access2rohit in #3005
- Upgrade vllm to v0.17.0 by @xyang16 in #3006
New Contributors
- @AdnaneKhan made their first contribution in #2979
- @johnpoth made their first contribution in #2994
- @access2rohit made their first contribution in #3001
Full Changelog: v0.35.0...v0.36.0