|
|
--- |
|
|
framework: pytorch |
|
|
model: EfficientNet-B1 |
|
|
dataset: CIFAR10 (restructured to have random upright and upside-down samples, with labels {0:'up', 1:'down'} |
|
|
imgSize: 224x224x3 |
|
|
--- |
|
|
|
|
|
|
|
|
This repository contains model trained to predict orientation {0:'up', 1:'down'} of images. |
|
|
The model '.pkl' file contains a dictionary with the following keys: |
|
|
|
|
|
```python: |
|
|
'optimizer': optimizer state dictionary |
|
|
'scheduler': scheduler state dictionary |
|
|
'model': model state dictionary |
|
|
'epoch': checkpoint epoch |
|
|
``` |
|
|
|
|
|
The image size is 3x224x224. The model can be initialized as: |
|
|
|
|
|
```python: |
|
|
model = torchvision.models.efficientnet_b1(pretrained=True) |
|
|
|
|
|
in_features = model.classifier[1].in_features |
|
|
out_features = 2 |
|
|
|
|
|
model.classifier[1] = nn.Linear(in_features, out_features, bias=True) |
|
|
``` |
