added onnx model

export_to_onnx.py

@@ -0,0 +1,50 @@
+from transformers import AutoModel, AutoTokenizer
+from pathlib import Path
+import torch
+import sys
+
+try:
+    print("Loading tokenizer...")
+    model_name = "."  # local dir
+    tokenizer = AutoTokenizer.from_pretrained(model_name)
+    print("✓ Tokenizer loaded successfully")
+    
+    print("Loading model...")
+    model = AutoModel.from_pretrained(model_name, trust_remote_code=True)
+    print("✓ Model loaded successfully")
+    
+    print("Setting model to evaluation mode...")
+    model.eval()
+    print("✓ Model set to evaluation mode")
+
+    print("Tokenizing input text...")
+    inputs = tokenizer("Export this model to ONNX!", return_tensors="pt")
+    print("✓ Input tokenized successfully")
+
+    print("Exporting model to ONNX format...")
+    # Export ONNX
+    torch.onnx.export(
+        model,
+        (inputs["input_ids"], inputs["attention_mask"]),
+        "model.onnx",
+        input_names=["input_ids", "attention_mask"],
+        output_names=["last_hidden_state"],
+        dynamic_axes={
+            "input_ids": {0: "batch", 1: "seq"},
+            "attention_mask": {0: "batch", 1: "seq"},
+            "last_hidden_state": {0: "batch", 1: "seq"},
+        },
+        opset_version=14,
+    )
+    print("✓ Model exported to ONNX successfully")
+    print(f"✓ ONNX file saved as: model.onnx")
+
+except FileNotFoundError as e:
+    print(f"❌ Error: Model files not found in current directory: {e}")
+    sys.exit(1)
+except ImportError as e:
+    print(f"❌ Error: Failed to import required modules: {e}")
+    sys.exit(1)
+except Exception as e:
+    print(f"❌ Error during model export: {e}")
+    sys.exit(1)

model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2459b4649b9b26e9ec6a72b4fea0e1ecd25fe4b24dbefe3db6e1d7df191a844
+size 436268726

modeling.py

@@ -131,6 +131,13 @@ class ConstBERT(BertPreTrainedModel):
 
         return torch.nn.functional.normalize(Q, p=2, dim=2)
 
+    def forward(self, input_ids, attention_mask):
+        """
+        Forward method for ONNX export and PyTorch compatibility.
+        This simply calls the existing _query method, preserving all current model behavior.
+        """
+        return self._query(input_ids, attention_mask)
+
     def _doc(self, input_ids, attention_mask, keep_dims=True):
         assert keep_dims in [True, False, 'return_mask']