add CLI support ArgillaTrainer

[davidberenstein1957]

Is your feature request related to a problem? Please describe.
I want to use CLI!

Describe the solution you'd like

Usage: python -m argilla.tasks.training [OPTIONS]

  Creates a default training task using CLI commands.

Options:
  -n, --name TEXT              The name of the dataset to be used for
                               training.
  -f, --framework TEXT         Choose a framework: transformers, openai,
                               span_marker, setfit or spacy.
  -w, --workspace TEXT         The workspace to be used for training.
  -l, --limit TEXT             The number of record to be used.
  -q, --query TEXT             The query to be used.
  -m, --model TEXT             The modelname or path to be used for training.
  -t, --train-size FLOAT       The train split to be used.
  -s, --seed INTEGER           The random seed number.
  -d, --device INTEGER         The GPU id to be used for training.
  -o, --output-dir TEXT        Output directory for the saved model.
  --update-config-kwargs TEXT  Output directory for the saved model.
  --api-url TEXT               Api url to be used for training.
  --api-key TEXT               Api key to be used for training.
  --help                       Show this message and exit.

Describe alternatives you've considered
N.A.

Additional context
N.A.

[alvarobartt]

To add something else, I think we can map the CLI entry point to be argilla train instead of argilla.tasks.training. And I'd suggest the usage of Typer.

[davidberenstein1957]

I agree.