ss

app.py

@@ -33,6 +33,21 @@ EMBEDDING_MODEL_NAME = "nvidia/NV-Embed-v2"
 print(f"Loading embedding model {EMBEDDING_MODEL_NAME}...")
 
 try:
+    # Create a simple Replicate class since it's missing
+    class Replicate(torch.nn.Module):
+        def __init__(self):
+            super().__init__()
+        
+        def forward(self, x):
+            return x
+    
+    # Add to global namespace to make it available
+    import sys
+    sys.modules['__main__'].Replicate = Replicate
+    
+    # Also add to globals
+    globals()['Replicate'] = Replicate
+    
     # Load embedding model and tokenizer
     global_embedding_tokenizer = AutoTokenizer.from_pretrained(EMBEDDING_MODEL_NAME, trust_remote_code=True)
     global_embedding_model = AutoModel.from_pretrained(EMBEDDING_MODEL_NAME, trust_remote_code=True, device_map="auto")

explanation_generator.py

@@ -17,6 +17,21 @@ print("Loading Qwen/QwQ-32B model with 4-bit quantization...")
 QWQ_MODEL_NAME = "Qwen/QwQ-32B"
 
 try:
+    # Create a simple Replicate class since it's missing
+    class Replicate(torch.nn.Module):
+        def __init__(self):
+            super().__init__()
+        
+        def forward(self, x):
+            return x
+    
+    # Add to global namespace to make it available
+    import sys
+    sys.modules['__main__'].Replicate = Replicate
+    
+    # Also add to globals
+    globals()['Replicate'] = Replicate
+    
     # Configure 4-bit quantization for better performance
     quantization_config = BitsAndBytesConfig(
         load_in_4bit=True,

requirements.txt

@@ -18,4 +18,3 @@ huggingface-hub>=0.30.0,<1.0
 einops
 bitsandbytes>=0.41.0
 accelerate>=0.23.0
-replicate==0.17.0