- Performed different learning procedures on the STL10 dataset - supervised learning, semi-supervised learning and self-supervised learning
- Used ResNet-50 Architecture and got validation accuracy of 68.7
- Used Pseudo-Labeling method using the same encoder architecture as in supervised learning
Model | Supervised Validation Accuracy | Semi-Supervised Validation Accuracy | Change in Accuracy |
---|---|---|---|
CNN Model | 59.4 | 64.62 | 5.08 |
ResNet-50 Model | 68.73 | 72 | 3.27 |
- For this I used the SimClr framework for contrastive learning and get a valiation accuracy of 53.30%
- I tried to implement semi-supervised tasks using SimClr and augment images using AutoAugment method. The operations we will be using are shearing, translating, rotation, auto_contrasting, brightness, sharpness, cutout, etc., and the policies for each augmentation are selected randomly and applied in our dataset for producing image augmentations