| Component | Points |
|---|---|
| 1. Implementation (35 points) | |
| Model Architecture Design | |
| • Appropriate model size for device constraints | |
| • Efficient layer organization | |
| • Memory optimization considerations | 10 |
| Model Optimization | |
| • Quantization implementation | |
| • Pruning techniques applied | |
| • Model compression methods | 15 |
| Device Integration | |
| • Successful deployment on target device | |
| • Hardware-specific optimizations | |
| • Error handling implementation | 10 |
| 2. Performance Metrics (25 points) | |
| Inference Speed | |
| • Latency measurements | |
| • Throughput analysis | 10 |
| Resource Utilization | |
| • Memory usage monitoring | |
| • CPU/GPU efficiency | |
| • Battery impact assessment | 8 |
| Accuracy Metrics (Initial Accuracy of the model will not be evaluated) | |
| • Model accuracy comparison | |
| • Performance degradation analysis | 7 |
| 3. Final presentation and Video (30 points) | |
| Technical Documentation done in GITHUB | |
| • Architecture diagrams | |
| • Implementation details | |
| • Setup instructions and source code | 10 |
| Performance Analysis | |
| • Benchmark results | |
| • Optimization decisions explained | |
| • Trade-off analysis | 5 |
| Demo Video | 15 |
| 4. Midterm (10 points) | |
| Progress Report | |
| • Clear timeline of completed work | |
| • Challenges encountered and solutions | |
| • Updated project scope | 2 |
| Technical Demonstration | |
| • Prototype demonstration | |
| • Initial performance metrics | |
| • Implementation progress | 4 |
| Future Plans | |
| • Remaining milestones | |
| • Risk assessment | |
| • Mitigation strategies | 2 |
| Presentation Quality | |
| • Clear communication | |
| • Professional slides | |
| • Q&A handling | 2 |
Total Points: 100