VLLM example code for serving SmolDocling or other VLM models

[TheMrguiller]

Question

Hi,

I have seen in other closed issues that the Docling team is working on incorporating VLLM serving into the image description extraction process. Is using this service, or a similar one, being considered for serving SmolDocling or other VLLMs directly, in order to enable not only image extraction but also text extraction for tables and others?