Template for LLM-based application built with Streamlit and OpenAI API (gpt-3.5-turbo
model).
This template was created for the Summarising Best Practices for Prompt Engineering article. Check it out for more details about prompt engineering techniques used to design prompts for LLMs.
Install streamlit
and openai
libraries by running pip install -r requirements.txt
.
Configure your OpenAI key by following instructions on the introduction to the OpenAI API page. See Generating OpenAI Key article section for more details.
With configured OpenAI key, the app extracts job title, company name, key skills, job summary and responsibilities from the provided job description. Prompt v5 is used - zero-shot with CoT for two-step summarization, output in JSON.
Note that without OpenAI key specified the app just waits two seconds and copies the input prompt to the output, appending "OpenAI key is not valid" note at the start.