Fix hugging face prompt pipeline driver + docs

[dylanholmes]

Changes:

• Edit example in Hugging Face Pipeline prompt driver docs to use HuggingFacePipelinePromptDriver instead of HuggingFacePromptDriver and remove api_token since it is not needed by HuggingFacePipelinePromptDriver.
• Add max_tokens field with a default value to HuggingFacePipelinePromptDriver in order to get it to work with the example. This is the same default provided in HuggingFacePromptDriver.
• Add a new extra drivers-prompt-huggingface-pipeline, which includes pytorch (torch) as a dependency to enable running models locally. Note that hugging face allows using pytorch or tensorflow, and I picked pytorch without any reasons in particular.
• Change Hugging Face Pipeline example's model choice to TinyLlama/TinyLlama-1.1B-Chat-v0.6 so that is small enough to run in the github action test runner as a part of the docs integration tests. (Plus it's nicer for users who want to test out the functionality faster locally)
• Add a couple of disclaimers about what models work with the HuggingFacePromptDriver.

Manual Testing:

• Successfully ran the updated Hugging Face Pipeline prompt driver example locally.

---

📚 Documentation preview 📚: https://griptape--796.org.readthedocs.build//796/

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

📢 Thoughts on this report? Let us know!

[collindutter]

Can you explain a bit why we need this now?

[dylanholmes]

Without this, running the example provided in the docs, I get:

/Users/dylan/Documents/boom/.venv/lib/python3.12/site-packages/stringcase.py:247: SyntaxWarning: invalid escape sequence '\W'
  return re.sub("\W+", "", string)
None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
[05/21/24 19:26:43] INFO     PromptTask ed3273d35309405e87eaacc7bb3360ee
                             Input: Hello Girafatron, what is your favorite animal?
WARNING:root:<RetryCallState 6048187696: attempt #1; slept for 0.0; last result: failed (RuntimeError At least one of TensorFlow 2.0 or PyTorch should be installed. To install TensorFlow 2.0, read the instructions at https://www.tensorflow.org/install/ To install PyTorch, read the instructions at https://pytorch.org/.)>
WARNING:root:<RetryCallState 6048187696: attempt #2; slept for 2.0; last result: failed (RuntimeError At least one of TensorFlow 2.0 or PyTorch should be installed. To install TensorFlow 2.0, read the instructions at https://www.tensorflow.org/install/ To install PyTorch, read the instructions at https://pytorch.org/.)>
WARNING:root:<RetryCallState 6048187696: attempt #3; slept for 4.0; last result: failed (RuntimeError At least one of TensorFlow 2.0 or PyTorch should be installed. To install TensorFlow 2.0, read the instructions at https://www.tensorflow.org/install/ To install PyTorch, read the instructions at https://pytorch.org/.)>

I was surprised to see this as well, and I can't explain why now. Though it does makes sense to need an ML framework for local ML inferencing.