modify the pipeline

app.py

@@ -31,11 +31,11 @@ import tempfile
 # Prompt template
 template = """Context: {context}
 
-Question: {question}
+Question: {query}
 
-Answer: Based on the provided context, """
+Answer:"""
+QA_PROMPT = PromptTemplate(template=template, input_variables=["query", "context"])
 
-QA_PROMPT = PromptTemplate(template=template, input_variables=["question", "context"])
 
 # Load Phi-2 model from hugging face hub
 model_id = "microsoft/phi-2"
@@ -148,18 +148,16 @@ def generate(question, answer, text_file, max_new_tokens):
         )
         phi2_pipeline = pipeline(
             "text-generation",
-            tokenizer=tokenizer,
             model=model,
+            tokenizer=tokenizer,
             max_new_tokens=max_new_tokens,
+            do_sample=False,               # ← greedy
             pad_token_id=tokenizer.eos_token_id,
             eos_token_id=tokenizer.eos_token_id,
-            do_sample=True,
-            temperature=0.7,
-            top_p=0.9,
-            repetition_penalty=1.1,
             streamer=streamer,
         )
 
+
         hf_model = HuggingFacePipeline(pipeline=phi2_pipeline)
         qa_chain, vectorstore = get_retrieval_qa_chain(text_file, hf_model)