Add customizable pytorch-huggingface VLM model fine-tuning pipeline example

[juk1329]

In the general_vlm_pytorch_huggingface branch, I added a fine-tuning and evaluation pipeline for VLM models. Previously, there was only a pipeline for language models that handled text data. With this VLM pipeline, it is now possible to work with vision data as well.

Similar to the language model pipeline, the VLM pipeline runs within the FastTrack environment, and all paths are properly redirected to FastTrack’s vFolder.

Since VLM models require more user configuration, the setup now involves four YAML files. Details can be found in the README.md.

If you have time, I would greatly appreciate any feedback, questions, or suggestions regarding the code or commit history.

Thank you very much for your time and review.