LlamaEdge 0.14.4Compare github-actions released this 20 Sep 06:28 · 1 commit to main since this release 0.14.4 1eb3fdd Major changes:chat-prompts crate New prompt types: nemotron-tool, nemotron-chat, and qwen-2.5-coder endpoints crate Improve the default values of TranslationRequest ...
Added new chat modules: Introduced falcon, megrez, and qwen modules along with their respective prompt types. Extended ChatPrompt enum: Included Phi4ChatPrompt, FalconChatPrompt, MegrezPrompt, and Qwen2vlPrompt. Updated PromptTemplateType mappings: Added new mappings for DeepseekChat3, Phi4Chat, ...
An error module with specific error types is defined, followed by an implementation for the Qwen25CoderInstructPrompt structure. The function build() checks if the messages vector is not empty, retrieves the last message from the vector, and handles it based on its type before returning the fi...
Use + to add a model, - to hide a model, and model_name=display_name to customize the display name of a model, separated by commas. qwen-7b-chat,+glm-6b,-gpt-3.5-turboNote The complete list of environment variables can be found in the 📘 Environment Variables...