remove minicpm3 for incompatibility reasons

utils/models.py

@@ -19,7 +19,7 @@ models = {
     #"Cogito-v1-preview-llama-3b": "deepcogito/cogito-v1-preview-llama-3b",
     "IBM Granite-3.3-2b-instruct": "ibm-granite/granite-3.3-2b-instruct",
     #"Bitnet-b1.58-2B4T": "microsoft/bitnet-b1.58-2B-4T",
-    "MiniCPM3-RAG-LoRA": "openbmb/MiniCPM3-RAG-LoRA",
+    #"MiniCPM3-RAG-LoRA": "openbmb/MiniCPM3-RAG-LoRA",
     "Qwen3-0.6b": "qwen/qwen3-0.6b",
     "Qwen3-1.7b": "qwen/qwen3-1.7b",
     "Qwen3-4b": "qwen/qwen3-4b",
@@ -160,7 +160,7 @@ def run_inference(model_name, context, question):
         if generation_interrupt.is_set():
             return ""
 
-        outputs = pipe(text_input, **generation_kwargs)
+        outputs = pipe(formatted, **generation_kwargs)
         result = outputs[0]['generated_text'][-1]['content']
 
     except Exception as e: