From 0c98c50a90a82b50de3729729c215a344f3c3c51 Mon Sep 17 00:00:00 2001 From: Raynor Chavez Date: Mon, 21 Oct 2024 16:16:32 +1100 Subject: [PATCH] 2.12.3, 2.12.4, 2.12.5 release notes --- RELEASE.md | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) diff --git a/RELEASE.md b/RELEASE.md index d30e57a22..c6d120f36 100644 --- a/RELEASE.md +++ b/RELEASE.md @@ -1,3 +1,18 @@ +# Release 2.12.5 + +## Bug fixes and minor changes +- Fix a bug where authentication is not correctly passed when loading a private model. Marqo can now load custom models from private repositories correctly ([#1010](https://github.com/marqo-ai/marqo/pull/1010)). + +# Release 2.12.4 + +## Bug fixes and minor changes +- Fix inference time regression from 2.11 to 2.12 where inference time increased unexpectedly ([#1005](https://github.com/marqo-ai/marqo/pull/1005)). + +# Release 2.12.3 + +## Bug fixes and minor changes +- Fix bug where users upgrading from indexes created with Marqo 2.11 to 2.12 would encounter an error when using the `get_settings` API endpoint ([#1003](https://github.com/marqo-ai/marqo/pull/1003)). + # Release 2.12.2 ## Bug fixes and minor changes @@ -10,7 +25,6 @@ - Fix a bug where when `treatUrlsAndPointersAsImages` is unset and `treatUrlsAndPointersAsMedia` is set, Marqo returns an error where `treatUrlsAndPointersAsImages` cannot be `False` when `treatUrlsAndPointersAsMedia` is `True` ([#971](https://github.com/marqo-ai/marqo/commit/a0084a86d5cf797616a1f8e185eba87417edbc15)) - Add new video-audio model `LanguageBind/Video_V1.5_FT_Audio_FT` to the model registry ([#971](https://github.com/marqo-ai/marqo/commit/a0084a86d5cf797616a1f8e185eba87417edbc15)). - # Release 2.12.0 ## New features @@ -18,7 +32,6 @@ - Load OpenCLIP models from HuggingFace Hub ([#939](https://github.com/marqo-ai/marqo/pull/939)). Support loading OpenCLIP models directly from HuggingFace by providing a model name with the hf-hub: prefix. This simplifies model integration and expands your options. - Load custom OpenCLIP checkpoints with different image preprocessors ([#939](https://github.com/marqo-ai/marqo/pull/939)). Allow loading a custom OpenCLIP checkpoint with a different image preprocessor by providing imagePreprocessor in the model properties. This offers greater flexibility in model selection and customization. - ## Bug fixes and minor changes - Fix tokenizer loading for custom OpenCLIP checkpoints ([#939](https://github.com/marqo-ai/marqo/pull/939)). The correct tokenizer is now applied when custom OpenCLIP model checkpoints are loaded. - Improve error handling for image_pointer fields in structured indexes ([#944](https://github.com/marqo-ai/marqo/pull/944)). Structured indexes now have targeted error reporting for non-image content in image_pointer fields. This improvement prevents batch failures and provides clearer feedback to users.