OpenAI GPT-2 based Python Script(s) To Train/Re-Train Custom Models and Generate Text from Prompts

The script(s) must utilise the OpenAI GPT-2 model and be simple to run, in order to generate text based on prompts. Most importantly, model must be re-trainable using large text and/or PDF files. The required script(s) would automatically make the text contained within multiple .txt and .pdf files (contained in a folder) usable for model training. No external editing or trimming of files should be needed for the intended script(s) to run. Training should be scale-able, and have the ability to be paused or resumed at any moment. The model should be able to generate text based on prompts just like the 1.5B trained OpenAI GPT-2 model. There should be scope to train multiple models, load them easily and generate custom prompts based on specific models. The script(s) must work in Windows 10 and require minimal coding experience and/or code typing. All script(s) must come with detailed instructions on how to operate.