Fixed deprecated torch_dtype argument from HuggingFace transformers: replaced with dtype to avoid warning

[RishabSA]

This PR resolves the warning from huggingface deprecation of the torch_dtype argument when loading a model from HuggingFace transformers. Replaced it with dtype whenever loading a model from transformers. Kept torch_dtype in kwargs since it automatically gets assigned to dtype when loading from TransformerLens.

Fixes #1093

Before:

hf_model = AutoModelForCausalLM.from_pretrained(
    official_model_name,
    torch_dtype=dtype,
    token=huggingface_token if len(huggingface_token) > 0 else None,
    **kwargs,
)

After:

hf_model = AutoModelForCausalLM.from_pretrained(
    official_model_name,
    dtype=dtype,
    token=huggingface_token if len(huggingface_token) > 0 else None,
    **kwargs,
)

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)

Checklist:

• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• I have not rewritten tests relating to key interfaces which would affect backward compatibility