Add types and MLX compatibility for cli generation

[nathanrchn]

1. Type Annotations

   • Added type hints throughout the codebase
   • Improves code readability and helps developers understand the library's structure and API

2. MLX Compatibility

   • Integrated MLX support for faster generation
   • New CLI argument: --use-mlx to enable MLX acceleration
   • Note: Quantization arguments are not yet compatible with MLX (--use_4bit and --use_8bit)
   • Note: I am still waiting to be merged on the mlx_lm repo, so the url to install mlx_lm will (maybe) change.

Usage Example

To use MLX acceleration:

transformers-cfg-cli generate \
    -m "mlx-community/Phi-3-mini-4k-instruct-4bit" \
    -g "examples/grammars/json.ebnf" \
    -p "This is a valid json string for http request:" \
    --use_mlx \
    --max_new_tokens 60 \
    --repetition_penalty 1.1
# {"name":"John","age":30,"car":null}