Update model with improved regularization and data augmentation
Browse files- README.md +2 -2
- confusion_matrix.png +0 -0
- per_class_metrics.csv +6 -6
- resnet18_beans.pth +1 -1
README.md
CHANGED
|
@@ -30,7 +30,7 @@ This model was trained in Google Colab using a T4 GPU and tracked with MLflow.
|
|
| 30 |
- Bean Rust
|
| 31 |
- Angular Leaf Spot
|
| 32 |
|
| 33 |
-
**Validation Accuracy:** 0.
|
| 34 |
|
| 35 |
## Training Configuration
|
| 36 |
|
|
@@ -46,7 +46,7 @@ This model was trained in Google Colab using a T4 GPU and tracked with MLflow.
|
|
| 46 |
- Epochs: 10
|
| 47 |
- Batch Size: 32
|
| 48 |
- Weight Decay: 0.0001
|
| 49 |
-
- Dropout: 0.
|
| 50 |
- Optimizer: Adam
|
| 51 |
|
| 52 |
## Artifacts
|
|
|
|
| 30 |
- Bean Rust
|
| 31 |
- Angular Leaf Spot
|
| 32 |
|
| 33 |
+
**Validation Accuracy:** 0.9398
|
| 34 |
|
| 35 |
## Training Configuration
|
| 36 |
|
|
|
|
| 46 |
- Epochs: 10
|
| 47 |
- Batch Size: 32
|
| 48 |
- Weight Decay: 0.0001
|
| 49 |
+
- Dropout: 0.3
|
| 50 |
- Optimizer: Adam
|
| 51 |
|
| 52 |
## Artifacts
|
confusion_matrix.png
CHANGED
|
|
per_class_metrics.csv
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
,precision,recall,f1-score,support
|
| 2 |
-
Healthy,0
|
| 3 |
-
Bean Rust,0.
|
| 4 |
-
Angular Leaf Spot,0
|
| 5 |
-
accuracy,0.
|
| 6 |
-
macro avg,0.
|
| 7 |
-
weighted avg,0.
|
|
|
|
| 1 |
,precision,recall,f1-score,support
|
| 2 |
+
Healthy,1.0,0.8863636363636364,0.9397590361445783,44.0
|
| 3 |
+
Bean Rust,0.8490566037735849,1.0,0.9183673469387755,45.0
|
| 4 |
+
Angular Leaf Spot,1.0,0.9318181818181818,0.9647058823529412,44.0
|
| 5 |
+
accuracy,0.9398496240601504,0.9398496240601504,0.9398496240601504,0.9398496240601504
|
| 6 |
+
macro avg,0.949685534591195,0.9393939393939393,0.9409440884787651,133.0
|
| 7 |
+
weighted avg,0.948928926088807,0.9398496240601504,0.940774338542374,133.0
|
resnet18_beans.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 44792395
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bdaaee60924b2d69d08666ce33f3b600e5a025a2cb41588aa24ec7b52127928f
|
| 3 |
size 44792395
|