This is a guided series of tutorials/notebooks that can be taken as a reference or course to build a PDF to Podcast workflow.
You will also learn from the experiments of using Text to Speech Models.
It assumes zero knowledge of LLMs, prompting and audio models, everything is covered in their respective notebooks.
P. S: This is Meta's Open Source response to Google’s NotebookLM, which I posted about earlier in Sept, link below.