Run python3 dev/update_changelog.py --prev-version 2.1...

Signed-off-by: mlflow-automation <[email protected]>

CHANGELOG.md

@@ -1,5 +1,36 @@
 # CHANGELOG
 
+## 2.12.0 (2024-03-14)
+
+MLflow 2.12.0 includes several major features and improvements
+
+Features:
+
+- [Models] Reapply "Implement promptflow model flavor (#10104)" (#10903) (#11311, @brynn-code)
+- [Models] Copy Model Metadata into subdirectory within artifact store for UC sharing (#11357, @WeichenXu123)
+- [Deployments] Implement chat & chat streaming for Anthropic in Deployments (#11195, @gabrielfu)
+- [Models] Support Map type (#11265, @WeichenXu123)
+- [] Add UCVolumeDatasetSource (#11301, @chenmoneygithub)
+
+Bug fixes:
+
+- [Models] Support transformers models with custom code (models requiring `trust_remote_code=True`) (#11412, @daniellok-db)
+- [Model Registry] Fix Local File Read/Path Traversal bypass (#11376, @WeichenXu123)
+- [] Update Keras autologging iteration input handling (#11394, @WeichenXu123)
+- [Build] Fix keras autologging train iteration dataset (#11383, @WeichenXu123)
+- [] Make async logging faster (#11346, @chenmoneygithub)
+- [UI] Add Commit Hash back to UI (#11324, @daniellok-db)
+- [] Explicitly import vectorstores and embeddings in databricks_dependencies (#11334, @daniellok-db)
+
+Documentation updates:
+
+- [Docs] [Doc] Add promptflow model flavor documentation (#11385, @brynn-code)
+- [Docs] Add doc for sentence-transformers OpenAI-compatible API (#11373, @es94129)
+
+Small bug fixes and documentation updates:
+
+#11419, #11416, #11418, #11417, #11415, #11408, #11336, #11335, @harupy; #11410, @brynn-code; #11389, #11378, #11303, @B-Step62; #11396, #11392, #11386, #11380, #11381, #11343, @WeichenXu123; #11370, @sunishsheth2009; #11379, #11304, @zhouyou9505; #11321, #11323, #11322, @michael-berk; #11358, @serena-ruan; #11375, #11345, @edwardfeng-db; #11369, @chenmoneygithub; #11355, #11354, #11351, #11349, #11339, #11338, @daniellok-db; #11333, @cdancette
+
 ## 2.11.1 (2024-03-06)
 
 MLflow 2.11.1 is a patch release, containing fixes for some Databricks integrations and other various issues.