lemonade-sdk/lemonade v7.0.3

on GitHub

Headline

Add support for Vision Language Model (VLM) GGUFs in Lemonade Server's llamacpp backend (@danielholanda).

• Added Gemma-3-4b-it-GGUF and Qwen2.5-VL-7B-Instruct VLMs to the suggested models list.
• Try it by installing one of those VLMs, selecting the VLM in Open WebUI, and sending an image with your message.

Additional Changes

• Add Llama-xLAM-2-8b-fc-r-Hybrid, the best OGA+Hybrid model for tool calling, to the suggested models list (@danielholanda)
• Replace the old installation guide markdown files with a new webapp (@jeremyfowers)
• Fix the pkg_resources deprecation notice (@danielholanda)
• Enable randomness in repeated prompts in the llm-prompt tool (@amd-pworfolk)

Full Changelog: v7.0.2...v7.0.3