transformers

c54b69a1 - feat(quantizers): support non-weight param names in TorchAo safetensors loading (#46325)

feat(quantizers): support non-weight param names in TorchAo safetensors loading (#46325) * feat(quantizers): support non-weight param names in TorchAo safetensors loading - Add _discover_quantized_param_names() to dynamically extract parameter names from safetensors metadata instead of hardcoding 'weight' - Supports MoE models with params like gate_up_proj, down_proj - Add aten.normal_ compat patch for TorchAO tensor subclasses * Replace aten.normal_ patch with _is_hf_initialized; filter quantized params by tensor type in metadata * Apply ruff formatting