Provide a better code snippet for generators in the "view API" page

[abidlabs]

Say you have a Gradio app like this:

import gradio as gr

def test(x):
    for i in range(len(x)):
        yield x[:i]

gr.Interface(test, "textbox", "textbox").queue().launch()

The view API page is still going to suggest that you use the .predict() function to call this method via an API:

However, this only gives you the first iteration, which is not usually what you want. Instead, it would be better to show the .submit() usage here.

Inspired by this question: https://discuss.huggingface.co/t/gradio-api-streaming/45395

[freddyaboulton]

This should also be reflected in the view_api method of the clients! Inspired by this internal question where someone is using predict to stream text from a chatbot:

https://huggingface.slack.com/archives/C02QZLG8GMN/p1689585284045839