Use `dtype` instead of deprecated `torch_dtype` for transformers >= 4.56

In src/pruna/algorithms/hqq.py line 272, AutoModelForCausalLM.from_pretrained is called with torch_dtype=...:

model = AutoModelForCausalLM.from_pretrained(
    temp_dir, quantization_config=..., torch_dtype=torch.float16 if ... else ...
)

The torch_dtype keyword argument was deprecated in transformers 4.56 (PR #39782) and replaced by dtype. On transformers 4.56+ this call emits a DeprecationWarning, and the argument will be removed in a future release, breaking HQQ quantization.

Suggested fix: choose the keyword based on the installed transformers version with packaging.version:

import transformers
from packaging.version import Version

def _dtype_kwargs(dtype):
    """`dtype` keyword of `from_pretrained` exists since transformers 4.56 (PR #39782);
    older versions use `torch_dtype`."""
    if Version(transformers.__version__) >= Version("4.56"):
        return {"dtype": dtype}
    return {"torch_dtype": dtype}

model = AutoModelForCausalLM.from_pretrained(
    temp_dir, quantization_config=...,
    **_dtype_kwargs(torch.float16 if ... else ...),
)

This keeps compatibility with transformers < 4.56 and stops the deprecation warning on 4.56+.

Start in src/pruna/algorithms/hqq.py at line 272 and inspect the AutoModelForCausalLM.from_pretrained call and its current dtype handling. Check the installed transformers version comparison approach described in the issue, then verify that transformers before 4.56 still receive torch_dtype while 4.56+ receives dtype and no deprecation warning is emitted.