Back to Education
Coursework
EuroSAT transfer-learning benchmark
A 7-paradigm comparison on satellite imagery: LoRA matches full fine-tuning at 0.36% of the trainable parameters and 3.7× faster.
Coursework (EuroSAT_DL_Assignment); numbers read straight from committed notebook outputs.
Results
Every value links to the committed artifact that produces it.
| Metric | Value | Evidence |
|---|---|---|
| Best overall — EfficientNet-B0, full fine-tune | 98.37% | eurosat_benchmark.ipynb (cell 100) |
| LoRA — ViT-B/16, r = 8 | 98.22% | eurosat_benchmark.ipynb (cell 94) |
| Trainable params under LoRA | 0.36% (310,292 / 86,108,948) | eurosat_benchmark.ipynb (cell 91) |
| LoRA vs full fine-tune, wall-clock | 3.70× faster (3.70 vs 13.69 min) | eurosat_benchmark.ipynb (cell 94) |
What didn’t work
SimCLR burned 490 minutes of A100 time to reach 89.0% — roughly 133× the compute for a worse result than a short transfer-learning run.
Reported as a null, not hidden — the same honesty policy applies to every number on this site.