Changelog
New Features:
- Video File Upload on Playground: You can now upload video files and have a model interpret the video. This only works with certain Gemini models that support video.
Bug Fixes:
- Huggingface: Fixed multiple bugs with the huggingface model integration. It now supports tool calling for non-streaming cases.
- Gemini: Fixed issue with manually setting the assistant role and tool call result metrics.
- OllamaEmbedder: Fixed issue where no embeddings were returned.
What's Changed
- add autogen performance test by @allenli178 in #2223
- Better OpenAI invoke error handling by @dirkbrnd in #2244
- Add tests and fix huggingface by @manthanguptaa in #2235
- update-cookbooks-ag-2755 by @ysolanky in #2252
- medical_imaging: fix error of google.genai.types.Part.from_bytes() a… by @Prabesh1sharma in #2228
- Video upload AG-2760 by @anuragts in #2257
- fix: Get embeddings from the correct URI path by @mishramonalisha76 in #2271
- Added base_url property in AzureOpenAI for non-default URLs by @martimfasantos in #2273
- Fix gemini no function to call by @dirkbrnd in #2270
- gemini-assistant-message-role-fix-ag-2768 by @ysolanky in #2263
- gemini-bug-fix-ag-2777 by @ysolanky in #2277
- Ensure Python 3.8 compatibility by replacing dict[str, Any] with typing.Dict by @ItsRoy69 in #2275
New Contributors
- @allenli178 made their first contribution in #2223
- @Prabesh1sharma made their first contribution in #2228
- @martimfasantos made their first contribution in #2273
Full Changelog: v1.1.7...v1.1.8