Update app.py

app.py

@@ -61,10 +61,18 @@ qwen = HuggingFaceEndpoint(
     do_sample=False,
 )
 
+qwen2 = HuggingFaceEndpoint(
+    repo_id="Qwen/Qwen2-1.5B-Instruct",
+    task="text-generation",
+    max_new_tokens=150,
+    do_sample=False,
+)
+
 llm = prompt_qwen | qwen
 
 llm2 = prompt_llama | llama
 
+llm3 = prompt_qwen | qwen2
 # llm = prompt | CustomLLM(repo_id="Qwen/Qwen-VL-Chat", model_type='text-generation', api_token=API_TOKEN, max_new_tokens=150).bind(stop=['<|im_end|>'])
 
 
@@ -82,6 +90,10 @@ async def conversation(data : ConversationPost):
 async def conversation2(data : ConversationPost):
     return {"output":llm2.invoke({"question":data.question})}
 
+@app.post("/conversation3")
+async def conversation3(data : ConversationPost):
+    return {"output":llm3.invoke({"question":data.question})}
+
 
 @app.post("/inference")
 async def inference(data : InferencePost):
@@ -89,6 +101,8 @@ async def inference(data : InferencePost):
         out = llm2.invoke(data.question)
     elif data.with_template == 'qwen':
         out = llm.invoke(data.question)
+    elif data.with_template == 'qwen2':
+        out = llm3.invoke(data.question)
     else:
         out = llama.invoke(data.question)