| question,answer |
| What is GAIA?,GAIA is Hugging Face's open framework for building language agent applications. |
| Which tools does GAIA support?,"GAIA supports tools like DuckDuckGoSearch, Python, and Wikipedia." |
| How do you use tools in GAIA?,By passing them to the AgentExecutor or using the with statement in Python. |
| What is the core of a GAIA agent?,"A PromptNode, an LLM-powered component that processes instructions." |
| What file defines an agent in the final assignment?,app.py |
| How many tools must be used in the final exam?,At least two. |
| Which LLM is commonly used in GAIA examples?,mistralai/Mistral-7B-Instruct-v0.2 |
| What should you include in your app for the certificate?,A callable Gradio interface that returns answers. |
| What does the scoring API expect?,"20 answers in order, submitted as a list." |
| Where do you submit your answers for scoring?,https://agents-course-unit4-scoring.hf.space/api/score |
| What is the format of the answers submission?,A JSON object with the key 'answers' and a list of 20 answers. |
| What does the Unit 4 exam test?,Your ability to build and submit a functional GAIA agent. |
| What Python package is required for the interface?,gradio |
| What is the purpose of PromptNode in GAIA?,To route and handle prompts using an LLM. |
| What kind of agent does the exam require?,A language agent built using the GAIA framework. |
| How do you get the certificate?,By answering at least 6 of 20 questions correctly. |
| What should your Gradio UI include?,A button to trigger answer generation and a result display. |
| What happens if the scoring API is unavailable?,The script should handle the error gracefully. |
| How are answers matched in fallback?,Using fuzzy matching with difflib. |
| What does the fallback model do?,Tries to guess answers using a question-answering pipeline. |
|
|
