fix dtype issue

vibevoice_nodes.py

@@ -14,6 +14,13 @@ import comfy.model_patcher
 from comfy.utils import ProgressBar
 from comfy.model_management import throw_exception_if_processing_interrupted
 
+# Import transformers and packaging to handle different library versions.
+import transformers
+from packaging import version
+
+_transformers_version = version.parse(transformers.__version__)
+_DTYPE_ARG_SUPPORTED = _transformers_version >= version.parse("4.56.0")
+
 from transformers import set_seed, AutoTokenizer, BitsAndBytesConfig
 from .vibevoice.modular.modeling_vibevoice_inference import VibeVoiceForConditionalGenerationInference
 from .vibevoice.processor.vibevoice_processor import VibeVoiceProcessor
@@ -229,12 +236,21 @@ class VibeVoiceLoader:
         
         try:
             logger.info(f"Loading model with dtype: {final_load_dtype} and attention: '{attn_implementation_for_load}'")
+            # Build a dictionary of keyword arguments for from_pretrained.
+            from_pretrained_kwargs = {
+                "attn_implementation": attn_implementation_for_load,
+                "device_map": "auto" if quant_config else device,
+                "quantization_config": quant_config,
+            }
+
+            # Use the correct dtype argument based on the transformers version.
+            if _DTYPE_ARG_SUPPORTED:
+                from_pretrained_kwargs['dtype'] = final_load_dtype
+            else:
+                from_pretrained_kwargs['torch_dtype'] = final_load_dtype
             model = VibeVoiceForConditionalGenerationInference.from_pretrained(
                 model_path,
-                dtype=final_load_dtype,
-                attn_implementation=attn_implementation_for_load,
-                device_map="auto" if quant_config else device,
-                quantization_config=quant_config,
+                **from_pretrained_kwargs
             )
             
             if attention_mode == "sage":