|
|
| ╔══════════════════════════════════════════════════════════════╗ |
| ║ DAVID MODEL SUMMARY ║ |
| ╠══════════════════════════════════════════════════════════════╣ |
| ║ ║ |
| ║ 🎯 VALIDATION ACCURACY: 84.66% ║ |
| ║ ║ |
| ╚══════════════════════════════════════════════════════════════╝ |
|
|
| MODEL: David-partial_shared-deep_efficiency |
| RUN ID: 20251012_161107 |
| BEST EPOCH: 10/10 |
|
|
| ═══════════════════════════════════════════════════════════════ |
|
|
| 📊 PERFORMANCE BREAKDOWN |
|
|
| Final Training Accuracy: 95.10% |
| Best Validation Accuracy: 84.66% |
|
|
| Per-Scale Accuracies: |
| • Scale 384: 84.75% |
| • Scale 512: 84.75% |
| • Scale 768: 84.66% |
| • Scale 1024: 84.38% |
| • Scale 1280: 84.36% |
| • Scale 1536: 84.26% |
| • Scale 1792: 84.24% |
| • Scale 2048: 84.26% |
|
|
| ═══════════════════════════════════════════════════════════════ |
|
|
| 🏗️ ARCHITECTURE |
|
|
| Preset: clip_vit_bigg14 |
| Sharing Mode: partial_shared |
| Fusion Mode: deep_efficiency |
| Scales: 8 scales - [384, 512, 768, 1024, 1280, 1536, 1792, 2048] |
| Feature Dim: 1280 |
| Parameters: 60,814,373 |
|
|
| ═══════════════════════════════════════════════════════════════ |
|
|
| 📈 TRAINING CURVE |
|
|
| Epoch | Train Acc | Val Acc | Learning Rate |
| ------|-----------|----------|-------------- |
| 1 | 81.03% | 83.71% | 9.76e-04 |
| 2 | 86.80% | 84.04% | 9.05e-04 |
| 3 | 88.32% | 84.37% | 7.94e-04 |
| 4 | 89.54% | 84.41% | 6.55e-04 |
| 5 | 90.74% | 84.58% | 5.01e-04 |
| 6 | 91.93% | 84.51% | 3.46e-04 |
| 7 | 93.03% | 84.59% | 2.07e-04 |
| 8 | 93.97% | 84.56% | 9.64e-05 |
| 9 | 94.68% | 84.61% | 2.54e-05 |
| 10 | 95.10% | 84.66% 👑 | 1.00e-03 |
|
|
| ═══════════════════════════════════════════════════════════════ |
|
|
| 📁 FILES |
|
|
| Best Model: best_model_acc84.66.safetensors |
| Config: david_config.json |
| Training Cfg: train_config.json |
| History: training_history.json |
|
|
| ═══════════════════════════════════════════════════════════════ |
|
|
| Generated: 2025-10-12 16:46:36 |
|
|
