Update to qwen:1.8b

app/rag.py

@@ -15,14 +15,15 @@ class ChatPDF:
     chain = None
 
     def __init__(self):
-        self.model = ChatOllama(model="llama3")
+        self.model = ChatOllama(model="qwen:1.8b")
         self.text_splitter = RecursiveCharacterTextSplitter(chunk_size=1024, chunk_overlap=64)
         self.prompt = PromptTemplate.from_template(
             """
-            <s> [INST] Use the context to answer the question.[/INST] </s> 
-            [INST] Question: {question} 
+            You are an assistant for question-answering tasks. Use the following pieces of context 
+            to answer the question. If you don't know the answer, just say that you don't know. 
+            Question: {question} 
             Context: {context} 
-            Answer: [/INST]
+            Answer: 
             """
         )
 

start_service.sh

@@ -7,7 +7,7 @@ ollama serve &
 sleep 5
 
 # Pull and run <YOUR_MODEL_NAME>
-ollama pull llama3
+ollama pull qwen:1.8b
 
 #
 fastapi run /code/app/main.py --port 7860