https://github.com/sharpiless/zskd-pytorch

Pytorch implement of Zero-shot Knowledge Distillation in Deep Neural Networks
https://github.com/sharpiless/zskd-pytorch

knowledge-distillation zero-data zero-shot-learning zskd

Last synced: about 2 months ago
JSON representation

Pytorch implement of Zero-shot Knowledge Distillation in Deep Neural Networks

• Host: GitHub
• URL: https://github.com/sharpiless/zskd-pytorch
• Owner: Sharpiless
• License: gpl-3.0
• Created: 2021-08-27T03:19:58.000Z (about 4 years ago)
• Default Branch: main
• Last Pushed: 2021-08-27T11:24:30.000Z (about 4 years ago)
• Last Synced: 2025-06-17T13:47:04.289Z (4 months ago)
• Topics: knowledge-distillation, zero-data, zero-shot-learning, zskd
• Language: Python
• Homepage:
• Size: 248 KB
• Stars: 6
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Zero-Shot Knowledge Distillation in Deep Networks Pytorch

|                   |        |         |               |

|-------------------|--------|---------|---------------|

| Model             | Method | Dataset | top1-accuracy |

| LeNet5-LeNet5Half | Paper  | Mnist   | 98.77         |

| LeNet5-LeNet5Half | Ours   | Mnist   | 96.98         |

# Usage

```bash

python main.py --dataset=mnist --lr=3.0 --t_train=False --num_sample=24000 --batch_size=100

```

# TO-DO

Hard to repruduce the results in paper. The paper author also does not reply.