ColeMurray/medium-show-and-tell-caption-generator ? reverse-engineered prompt

Build me a simple Python tool that can take one or more images and print out a caption for each one using a pre trained Show and Tell model.

I want it to work from the command line, with an option for the model file, the vocabulary file, and a folder or file pattern for input images. It should be easy to run either in Docker or in a local conda environment, and it should come with clear setup steps so I can download the model and test it on my own images.

Please make sure it reads the images from an imgs folder, loads the caption model, and outputs a sensible sentence for each image. If anything is missing, like the model files or vocabulary, give a helpful error message. If you need to check current docs or best practices online, feel free to do that.