I am new to pre-trained language models in natural language processing. Could anyone give a hint on where should I start, or the road maps to start?

I know there are quite many hands-on tutorials about deploying BERT or other models. But the problem is many of them are just shallow user cases which only modified a few parameters using apis from a certain libraries such as keras. I am a novice and I often find my still confused at the details and variational steps when deploying pre-trained language models on my projects.

Could anyone give a hint on what is a better roadmap for learning coding with pre-trained language models, including any resources, articles, or tutorials, etc.

I have read many articles on Medium. But most articles there seems only introduc the general concepts, rather than providing the real know-how when learning it.

I find these two courses very novice-friendly:

1. Hugging Face course https://huggingface.co/course/chapter1/1

2. Deep Lizard "Deep Learning Fundamentals - Classic Edition" https://deeplizard.com/learn/playlist/PLZbbT5o_s2xq7LwI2y8_QtvuXZedL6tQU