Add ludwig upload to push artifacts to HuggingFace Hub

[arnavgarg1]

This PR adds a new command called ludwig upload to push trained model artifacts to Huggingface Hub.

The API is very simple:

ludwig upload hf_hub --repo_id <repo_id> --model_path </path/to/saved/model>

When you run this command, it will check if you are already signed into Huggingface Hub. This can be done through huggingface-cli login in advance, otherwise it will make you follow the steps to sign in which looks something like this:

$ ludwig upload --service hf_hub -r arnavgrg/test -m /workspaces/trained_models/api_experiment_run

    _|    _|  _|    _|    _|_|_|    _|_|_|  _|_|_|  _|      _|    _|_|_|      _|_|_|_|    _|_|      _|_|_|  _|_|_|_|
    _|    _|  _|    _|  _|        _|          _|    _|_|    _|  _|            _|        _|    _|  _|        _|
    _|_|_|_|  _|    _|  _|  _|_|  _|  _|_|    _|    _|  _|  _|  _|  _|_|      _|_|_|    _|_|_|_|  _|        _|_|_|
    _|    _|  _|    _|  _|    _|  _|    _|    _|    _|    _|_|  _|    _|      _|        _|    _|  _|        _|
    _|    _|    _|_|      _|_|_|    _|_|_|  _|_|_|  _|      _|    _|_|_|      _|        _|    _|    _|_|_|  _|_|_|_|
    
    To login, `huggingface_hub` requires a token generated from https://huggingface.co/settings/tokens .
Token: 
Add token as git credential? (Y/n) Y
Token is valid (permission: write).
Your token has been saved in your configured git credential helpers (!f()).
Your token has been saved to /home/vscode/.cache/huggingface/token
Login successful

You can find your token at https://huggingface.co/settings/tokens.

Once you're signed in, this will create a new public model repository on Huggingface Hub called repo_id if it doesn't exist, and push the trained model weights and associated configs in model_path to repo_id.

There are some other optional parameters you can pass to ludwig upload:

1. private: Whether to create a private repository or a public repository.
2. commit_message: The summary / title / first line of the generated commit.
3. commit_description: The description of the generated commit.

[github-actions]

Unit Test Results

       6 files  ±       0         6 suites  ±0   1h 14m 58s ⏱️ + 11m 47s
2 775 tests +2 741  2 756 ✔️ +2 727  12 💤 +7  7 ❌ +7 
2 818 runs  +2 730  2 790 ✔️ +2 718  21 💤 +5  7 ❌ +7 

For more details on these failures, see this check.

Results for commit 1d11569. ± Comparison against base commit 62a5524.

♻️ This comment has been updated with latest results.

[tgaddair]

This is awesome!

I'm kind of torn on the naming. I think commit is somewhat counterintuitive to me. The HF Hub API itself uses the terminology "upload" which feels more clear. Maybe we can do that instead? I also think making it clear we're uploading to HF (as opposed to Predibase or MLFlow or whatever) is important.

So to that end I would recommend the following:

ludwig upload hf_hub ...

What do you think?

[tgaddair]

Does HF support non-HF models being uploaded to the hub?