Example of response generation with optional arguments

[a-wozniakowski]

The example is meant to extend the README, whereby it:

• uses a checkpoint from the mlx-community
• loads the model and tokenizer with load
• transforms a prompt into a chat template with the apply_chat_template method of the tokenizer
• passes optional arguments into generate, which show how to control the generation process