https://github.com/timcsy/cnn_cifar10
https://github.com/timcsy/cnn_cifar10
Last synced: 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/timcsy/cnn_cifar10
- Owner: timcsy
- Created: 2020-10-27T06:07:43.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2020-10-30T21:54:27.000Z (over 4 years ago)
- Last Synced: 2025-01-18T23:15:02.174Z (4 months ago)
- Language: Jupyter Notebook
- Size: 9.7 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Recognizing Cifar10 using VGG16 / All-CNN
## Requirements
- Python 3 64-bit
- Dependency
```
tkinter tensorflow matplotlib numpy
```
## Usage
run:
```
python cnn.py
```
## Result
## 心得
- Colab 可以用 GPU 加速！！
- 用 VGG16 訓練一開始，Accuracy 結果都趨近於0.1（隨便亂猜的意思）
- VGG16 因為會 Gradient vanish，所以要做 batch normalize
- 其實也有其他的 Model 可以實作，我也有用 All-CNN 來寫
## Reference
- [Convolutional Neural Network (CNN)](https://www.tensorflow.org/tutorials/images/cnn)
- [Densely Connected Convolutional Networks](https://arxiv.org/pdf/1608.06993.pdf)
- Not used
- Found the following reference
- [STRIVING FOR SIMPLICITY: THE ALL CONVOLUTIONAL NET](https://arxiv.org/pdf/1412.6806.pdf)
- Use the All-CNN-C Model
- [All-CNN](https://github.com/PAN001/All-CNN)
- [python - 使用子类模型时，model.summary()无法打印输出形状](https://www.coder.work/article/1258695)