[Python] How to remove the <think> tag in reasoning model DeepSeek responses

I'm using deepseek llm for my AI app, the responses will include its thinking process.

It looks something like this:

the code generate the response :

# Generate a new response if last message is not from assistant
if st.session_state.messages[-1]["role"] != "assistant":
with st.chat_message("assistant"):
with st.spinner("Thinking..."):
response = ask_question(qa, prompt)
print(response.content)
st.markdown(response.content)
message = {"role": "assistant", "content": response.content}
st.session_state.messages.append(message)


The responses of the AI assistant has content in tags.

The actual answer is provided after the tags

do you have any idea please how to remove the think tag please ?

Continue reading...