Implementation of CRD (Contrastive Representation Distillation)
python main.py --option "teacher"
python main.py --option "student"
python main.py --option "distill" --teacher_model [teacher_model]
python main.py --option [option] --test --prev_model [prev_model]
Model | Teacher | Student (No Distill) | Studnet (Distill) | ||||
---|---|---|---|---|---|---|---|
Beta | - | - | 0.7 | 0.8 | 0.9 | 1.0 | 1.1 |
Accuracy | 77.08% | 76% | 75.92% | 75.57% | 76.08% | 76.27% | 75.74% |
- Contrastive Representation Distillation (https://arxiv.org/abs/1910.10699)
- CIFAR-100 Dataset (https://www.cs.toronto.edu/~kriz/cifar.html)
- ResNet implementation (https://github.com/weiaicunzai/pytorch-cifar100/blob/master/models/resnet.py)