* support for sglang 0.3.0 and onevision, etc * update notebooks to point llama3 llava as default model
0.8.0
* replicating the experience of dbx model serving and have settings for testing to be as close as possible to databricks model serving * updated docs on how to use the testing fixtures
0.7.3
* fix defaults for testing local env
0.7.2
* environment reuse for local testing without spinning up new python using pyenv and pip
0.7.1
* added sanity tests for sglang * added better test fixtures for testing serving deployments before deploying them to databricks
0.7.0
* upgraded to vllm 0.6.0 closes https://github.com/stikkireddy/mlflow-extensions/issues/4 * fixes https://github.com/stikkireddy/mlflow-extensions/issues/24 * introduces a sanity test for vllm to test all supported ezdeploy models before releasing/upgrading new versions