0.11.0

0.11.0 (2024-11-06)

Features

• Support method HEAD for /api/ping too (5fd367b)
• Add docker compose example for ovai & ollama (04677c1)

0.10.0

0.10.0 (2024-11-02)

Features

• Remove support for legacy PaLM 2 models (3325c54)

BREAKING CHANGES

Models text-bison* chat-bison* and text-unicorn*
have been removed. They can't be used in new projects. And they'll be
disabled on 9th April, 2025.

0.9.0

0.9.0 (2024-11-02)

Features

• Recognise multimodalembedding model (242709f)
• Implement streaming mode (e3989da)

0.8.0

0.8.0 (2024-11-02)

Features

• Add new models and model versions released by Google this year (0aeeff1)

BREAKING CHANGES

Models chat-bison@001 and text-bison@001 were removed.
Google stopped supporting them.

0.7.1

0.7.1 (2024-06-09)

Bug Fixes

• Base the image on busybox (706c2fa)

0.7.0

0.7.0 (2024-06-09)

Features

• Recognise model gemini-1.0-pro-vision (d042aec)
• Support attaching images to generate and chat requests (a3d7f15)
• Recognise the model multimodalembedding@001 (5c3aabe)

Bug Fixes

• Use padded base64 for JWT parts (8d01c1c)
• Do not log preflighting as an error (de96929)

0.6.0

0.6.0 (2024-06-05)

Features

• Allow forcing either IPV4 or IPV6 network connections only (5c0a9e4)

Bug Fixes

• Allow scope and auth_uri empty in the account file (54dbb29)

0.5.0

0.5.0 (2024-06-04)

Features

• Add chat, generate and embeddings purposes to the model info (cbfe399)
• Recognise model text-unicorn (8e83e7e)

Bug Fixes

• Recognise names of all PalLM embedding models (b602248)

0.4.0

0.4.0 (2024-06-04)

Features

• Support /api/tags for listing available models (fa5b1f0)
• Support /api/show for inspecting model information (8a2c2bc)

Bug Fixes

• Propagate error message from requests proxied to ollama (d2f59f1)

0.3.0

0.3.0 (2024-06-03)

Features

• Allow forwarding requests to ollama for other than Vertex AI models (8078a19)