PopularMajor Project
Artificial Intelligence
Face Recognition Attendance System
Camera-based attendance with liveness detection and analytics
- Python
- OpenCV
- FaceNet
- Flask
- +2
₹4,900₹5,999
4.8/301 reviews
Artificial Intelligence · Capstone
Code Generation Model with LLMs
Fine-tunes a code model and scores it with pass@k on unit tests
5-6 weeks build
Overview
Adapts a small open code model to Python function completion and evaluates it the way the field does: by running the generated code against unit tests. The public HumanEval and MBPP benchmarks are executed inside a sandboxed container, and pass@k is computed for the base and fine-tuned models across several sampling temperatures.
What makes it stand up
- 01LoRA fine-tuning of a small open code model
- 02Sandboxed Docker execution of generated code with time limits
- 03Unbiased pass@k estimator implemented and explained
- 04HumanEval and MBPP results for base and fine-tuned models
- 05Completion API with a simple editor front end
Module breakdown
- 01
Data
Docstring-to-function pairs, filtering and deduplication.
- 02
Fine-tuning
PEFT LoRA training with a code-aware prompt format.
- 03
Sandbox
Isolated containers with CPU, memory and time caps.
- 04
Evaluation
pass@k over multiple samples and temperatures.
- 05
Serving
FastAPI completion endpoint with a browser editor.
After this, you will be able to
- Explain pass@k and why it beats text-matching metrics for code
- Defend your sandboxing choices for untrusted code
- Demo completion on a function stub written on the day
Technology stack
- Python
- PyTorch
- Transformers
- PEFT
- Docker
- FastAPI
- Level
- Capstone
- Domain
- Artificial Intelligence
- Build time
- 5-6 weeks
- Difficulty
You receive
- Complete, commented source code
- 60-90 page project report (IEEE format)
- Editable presentation deck
- Architecture, ER and UML diagrams
- Local setup and deployment guide
- Sample dataset or seed data
- Viva question bank with answers
- One 45-minute walkthrough call
- 30 days of doubt-clearing support