Description
Feature request
Currently, when I check phi-4-mini model with TGI and also phi-4-multimodal model, none of them can be loaded with TGI version 3.1.1. Phi-4-mini need transformer version 4.49.0 and phi-4-multimodal got the following error:
Unsupported model type phi4mm
Is there any possibility to support these two models?
Also is it possible to use TensorRT-LLM backend?
Motivation
Resolve the issues with phi-4-mini and phi-4-multimodal model to be deployed with TGI
Your contribution
Upgrade transformer to version 4.49.0 and add support for multimodal models