Update README.md
Browse files
README.md
CHANGED
|
@@ -47,13 +47,13 @@ The model we are finetuning, microsoft/swin-large-patch4-window12-384-in22k, was
|
|
| 47 |
|
| 48 |
### Finetuning Data
|
| 49 |
|
| 50 |
-
The training data consists of
|
| 51 |
We finetuned the model on 3533 samples of the labeled dataset we were given, stratified on the label.
|
| 52 |
|
| 53 |
|
| 54 |
#### Preprocessing
|
| 55 |
|
| 56 |
-
Data augmentation was applied to the training data in a custom Torch dataset class. Because of the size of the dataset images were not replaced but duplicated and augmented.
|
| 57 |
The only augmentations applied were HorizontalFlips and Rotations (10 degrees) to align with the relatively homogenous dataset.
|
| 58 |
|
| 59 |
#### Finetuning Hyperparameters
|
|
@@ -70,22 +70,10 @@ The only augmentations applied were HorizontalFlips and Rotations (10 degrees) t
|
|
| 70 |
|
| 71 |
|
| 72 |
## Evaluation
|
| 73 |
-
The evaluation data consists of
|
| 74 |
We evaluated the model on 393 samples of the labeled dataset we were given, stratified on the label.
|
| 75 |
### Testing Data, Factors & Metrics
|
| 76 |
|
| 77 |
#### Testing Data
|
| 78 |
|
| 79 |
-
The testing data consists of an
|
| 80 |
-
|
| 81 |
-
[More Information Needed]
|
| 82 |
-
|
| 83 |
-
#### Factors
|
| 84 |
-
|
| 85 |
-
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 86 |
-
|
| 87 |
-
[More Information Needed]
|
| 88 |
-
|
| 89 |
-
#### Metrics
|
| 90 |
-
|
| 91 |
-
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
|
|
|
| 47 |
|
| 48 |
### Finetuning Data
|
| 49 |
|
| 50 |
+
The training data consists of a subset of the cub-200-2011 dataset, https://paperswithcode.com/dataset/cub-200-2011.
|
| 51 |
We finetuned the model on 3533 samples of the labeled dataset we were given, stratified on the label.
|
| 52 |
|
| 53 |
|
| 54 |
#### Preprocessing
|
| 55 |
|
| 56 |
+
Data augmentation was applied to the training data in a custom Torch dataset class. Because of the size of the dataset, images were not replaced but were duplicated and augmented.
|
| 57 |
The only augmentations applied were HorizontalFlips and Rotations (10 degrees) to align with the relatively homogenous dataset.
|
| 58 |
|
| 59 |
#### Finetuning Hyperparameters
|
|
|
|
| 70 |
|
| 71 |
|
| 72 |
## Evaluation
|
| 73 |
+
The evaluation data consists of a subset of the cub-200-2011 dataset, https://paperswithcode.com/dataset/cub-200-2011
|
| 74 |
We evaluated the model on 393 samples of the labeled dataset we were given, stratified on the label.
|
| 75 |
### Testing Data, Factors & Metrics
|
| 76 |
|
| 77 |
#### Testing Data
|
| 78 |
|
| 79 |
+
The testing data consists of an unlabeled subset of the cub-200-2011 dataset, https://paperswithcode.com/dataset/cub-200-2011 of 4000 images.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|