shazaaad/agents-course-final-agent

🕵️‍♂️ Hugging Face Agents Course — Final Agent

A production-ready AI agent built with Hugging Face smolagents.

Piece	What it does
CodeAgent	Writes short Python actions to call tools
InferenceClientModel	Free Hub inference (Qwen/Qwen2.5-Coder-32B-Instruct)
Tools	Web search, visit webpage, calculator, text summarizer
Gradio UI	Chat tab + Unit 4 certificate evaluation / submit tab
submit_answers.py	Local CLI that submits scores without needing a Space

Note on model class names: older tutorials mentioned HuggingFaceHubServerModel / HfApiModel.
In current smolagents, that Hub backend is called InferenceClientModel.

Important (2026): Creating new Gradio Spaces on Hugging Face now requires a PRO plan (or an eligible older account for ZeroGPU).
If you cannot create a Space, use the local runner below — it still submits to the official Unit 4 leaderboard.

Public code: https://huggingface.co/shazaaad/agents-course-final-agent

---

What you need for the certificate

To earn the Certificate of Completion you must score ≥ 30% on the Unit 4 GAIA subset leaderboard.

1. Run the agent (Space or local submit_answers.py).
2. Submit answers to the scoring API.
3. Check your score on the student leaderboard.

Also complete Unit 1 and one use-case assignment as required by the course.

---

Set up your Hugging Face API token (HF_TOKEN)

A) On a Hugging Face Space (if you have PRO / eligible ZeroGPU)

1. Open your Space → Settings → Variables and secrets.
2. Add secret HF_TOKEN = your token from https://huggingface.co/settings/tokens.

B) On your computer (works without Spaces PRO)

1. Create a token at https://huggingface.co/settings/tokens.
2. Copy .env.example → .env and paste:

HF_TOKEN=hf_your_token_here

Or run: hf auth login

---

Run locally (4 steps)

1. Open a terminal in this project folder.
2. Create & activate a virtual environment:

python -m venv .venv

• Windows PowerShell: .venv\Scripts\Activate.ps1
• macOS / Linux: source .venv/bin/activate

3. Install dependencies:

pip install -r requirements.txt

4. Add your token, then either:

Chat UI

python app.py

Certificate submission (recommended if you have no Space)

python submit_answers.py

Open the local URL Gradio prints (usually http://127.0.0.1:7860) when using app.py.

---

Push to a Hugging Face Space (if your account allows Gradio)

1. Create a Space at https://huggingface.co/new-space (Gradio + CPU basic / ZeroGPU).
2. Upload app.py, requirements.txt, and README.md.
3. Add secret HF_TOKEN.
4. Open Certificate Evaluation → Log in → Submit.
5. Score ≥ 30% → claim your certificate.

Optional: push with CLI

hf upload YOUR_USERNAME/YOUR_SPACE_NAME . --repo-type space

---

Project files

hf-agents-course-final/
├── app.py              # Agent + Gradio UI (chat + certificate submit)
├── submit_answers.py   # Local leaderboard submission (no Space needed)
├── requirements.txt    # Python packages
├── README.md           # This file
├── .env.example        # Template for local HF_TOKEN
└── .gitignore

Good luck — you’ve got this! 🎓