Upgrade transformers to 5.5.3 and refactor hf_transformers_utils into subpackage (#21569)

python/pyproject_xpu.toml

@@ -27,7 +27,9 @@ dependencies = [
   "blobfile==3.0.0",
   "build",
   "compressed-tensors",
+  "addict",
   "datasets",
+  "easydict",
   "einops",
   "fastapi",
   "gguf",
@@ -60,8 +62,8 @@ dependencies = [
   "timm==1.0.16",
   "torchao==0.9.0+xpu",
   "tqdm",
-  "mistral_common>=1.9.0",
-  "transformers==5.3.0",
+  "mistral_common>=1.11.0",
+  "transformers==5.5.4",
   "uvicorn",
   "uvloop",
   # "xgrammar==0.1.24", , xgrammar depends on CUDA PyTorch and Triton only