[Doc]: NVFP4 minimal example code

📚 The doc issue

Hi, NVFP4 seems like a really good format, and may well be used for training as well as quantization in the future. However I couldn't find any documentation how this actually works, and also not in the code. There's only this blog post.

Is there a simple pytorch function that demonstrates how to fake quantize a tensor from FP32 to NVFP4?
A small example to show how it works, without requiring the full TensorRT-LLM or TransformerEngine dependency, would be very helpful.

Suggest a potential alternative/fix

No response

Before submitting a new issue...

• Make sure you already searched for relevant issues, and checked the documentation and examples for answers to frequently asked questions.