Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
# Training Logs - EfficientNet-B3
|
| 2 |
-
## Experiment:
|
| 3 |
|
| 4 |
### Model Architecture
|
| 5 |
- **Base Model**: EfficientNet-B3 (pretrained on ImageNet)
|
|
@@ -11,28 +11,28 @@
|
|
| 11 |
- **Learning Rate**: 0.0002
|
| 12 |
- **Optimizer**: AdamW (weight_decay=1e-4)
|
| 13 |
- **Scheduler**: CosineAnnealingWarmRestarts
|
| 14 |
-
- **Epochs Trained**:
|
| 15 |
|
| 16 |
### Best Results
|
| 17 |
-
- **Validation Harmonic Mean**: 0.
|
| 18 |
-
- **Validation Age RMSE**: 0.
|
| 19 |
-
- **Validation Gender F1**: 0.
|
| 20 |
-
- **Best Epoch**:
|
| 21 |
|
| 22 |
### Final Metrics (Last Epoch)
|
| 23 |
-
- **Val Loss**: 0.
|
| 24 |
-
- **Val Harmonic Mean**: 0.
|
| 25 |
-
- **Val Age RMSE**: 0.
|
| 26 |
-
- **Val Age MAE**: 0.
|
| 27 |
-
- **Val Gender F1**: 0.
|
| 28 |
|
| 29 |
### Files in this Dataset
|
| 30 |
-
- `pytorch_lightning/
|
| 31 |
-
- `summaries/
|
| 32 |
-
- `trackio/
|
| 33 |
|
| 34 |
### Next Steps
|
| 35 |
Use the best checkpoint for inference:
|
| 36 |
```python
|
| 37 |
-
checkpoint_path = "checkpoints/efficientnet/effnet-epoch=
|
| 38 |
```
|
|
|
|
| 1 |
# Training Logs - EfficientNet-B3
|
| 2 |
+
## Experiment: 20251111_031138
|
| 3 |
|
| 4 |
### Model Architecture
|
| 5 |
- **Base Model**: EfficientNet-B3 (pretrained on ImageNet)
|
|
|
|
| 11 |
- **Learning Rate**: 0.0002
|
| 12 |
- **Optimizer**: AdamW (weight_decay=1e-4)
|
| 13 |
- **Scheduler**: CosineAnnealingWarmRestarts
|
| 14 |
+
- **Epochs Trained**: 34 / 40
|
| 15 |
|
| 16 |
### Best Results
|
| 17 |
+
- **Validation Harmonic Mean**: 0.9699
|
| 18 |
+
- **Validation Age RMSE**: 0.1175 years
|
| 19 |
+
- **Validation Gender F1**: 0.9469
|
| 20 |
+
- **Best Epoch**: 12
|
| 21 |
|
| 22 |
### Final Metrics (Last Epoch)
|
| 23 |
+
- **Val Loss**: 0.2847
|
| 24 |
+
- **Val Harmonic Mean**: 0.9650
|
| 25 |
+
- **Val Age RMSE**: 0.1177
|
| 26 |
+
- **Val Age MAE**: 0.0875
|
| 27 |
+
- **Val Gender F1**: 0.9375
|
| 28 |
|
| 29 |
### Files in this Dataset
|
| 30 |
+
- `pytorch_lightning/metrics_20251111_031138.csv` - Complete training history
|
| 31 |
+
- `summaries/training_summary_20251111_031138.json` - Detailed metrics and config
|
| 32 |
+
- `trackio/trackio_logs_20251111_031138.zip` - TrackIO logs (if available)
|
| 33 |
|
| 34 |
### Next Steps
|
| 35 |
Use the best checkpoint for inference:
|
| 36 |
```python
|
| 37 |
+
checkpoint_path = "checkpoints/efficientnet/effnet-epoch=12-val_harmonic_mean=0.9699.ckpt"
|
| 38 |
```
|