diff --git a/example_trainer/grpo.py b/example_trainer/grpo.py
index cca23b33..b010e551 100644
--- a/example_trainer/grpo.py
+++ b/example_trainer/grpo.py
@@ -454,21 +454,46 @@ def _attach_to_vllm_shared_tensors(
         print("[Setup] Single-copy mode not available (no IPC handles exported)")
         return None
     
-    ipc_handles = bridge_config.get("ipc_handles", {})
-    if not ipc_handles:
+    ipc_handles_raw = bridge_config.get("ipc_handles", {})
+    if not ipc_handles_raw:
         print("[Setup] No IPC handles found in bridge config")
         return None
     
+    # Deserialize base64-encoded bytes back to bytes
+    import base64
+    
+    def deserialize_ipc_handles(handles):
+        result = {}
+        for k, v in handles.items():
+            if isinstance(v, dict):
+                if "_bytes_b64_" in v:
+                    result[k] = base64.b64decode(v["_bytes_b64_"])
+                else:
+                    result[k] = deserialize_ipc_handles(v)
+            else:
+                result[k] = v
+        return result
+    
+    ipc_handles = deserialize_ipc_handles(ipc_handles_raw)
+    
     print(f"[Setup] Attaching to vLLM's shared tensors ({len(ipc_handles)} tensors)...")
     print("[Setup] TRUE SINGLE-COPY MODE - No additional model memory!")
     
-    # Create model architecture (meta device - no memory allocation)
+    # Load model config (not weights) to get architecture
+    from transformers import AutoConfig
+    model_config = AutoConfig.from_pretrained(config.model_name)
+    
+    # Create empty model on meta device (no memory allocation)
     with torch.device('meta'):
-        model = AutoModelForCausalLM.from_pretrained(
-            config.model_name,
+        model = AutoModelForCausalLM.from_config(
+            model_config,
             torch_dtype=torch.bfloat16,
         )
     
+    # Get parameter names from the empty model
+    param_names = list(model.state_dict().keys())
+    print(f"[Setup] Model architecture has {len(param_names)} parameters")
+    
     # Map vLLM tensor names to HuggingFace model parameter names
     hf_state_dict = {}
     vllm_to_hf_mapping = _create_vllm_to_hf_mapping(model, ipc_handles)
@@ -482,7 +507,16 @@ def _attach_to_vllm_shared_tensors(
         
         try:
             # Reconstruct tensor from IPC handle
-            handle_bytes = bytes.fromhex(ipc_info["handle"])
+            # Handle can be bytes (deserialized from base64) or hex string
+            handle = ipc_info["handle"]
+            if isinstance(handle, bytes):
+                handle_bytes = handle
+            elif isinstance(handle, str):
+                handle_bytes = bytes.fromhex(handle)
+            else:
+                print(f"[Setup] Unknown handle type for {hf_name}: {type(handle)}")
+                continue
+                
             storage_size = ipc_info["storage_size"]
             device_index = ipc_info["device_index"]