| The First Column | The Second Column |
|---|---|
| Debugprint |  |
| DebugPysnooper |  |
| slim | |
| resnetidentitymapping |  |
| resnetbuildingblock |  |
| densenet | |
| densenet整体结构 |  |
| densenet整体结构 |  |
| densenetFeature Block |  |
| densenetDense Layer |  |
| densenetDense Block |  |
| densenetTransition Block |  |
| densenet循环 Dense Block和Transition |  |
| densenetClassificationBlock |  |
 | |
| inception | |
| label number 不均 |  |
| idea | 确定研究方向 |
| fc 的节点 |  |
| channel num_filters | channel are input_feature_map_number num_filters are output_feature_map_number |
| VGG16fully connected layer |  |
| VGG16fully connected layer |  |
| VGG16fully connected layer |  |
| 类别不均 |  |
| 类别不均 |  |
 | |
| 迁移学习图像尺寸 |  |
| learning rate decay |  |
| 各种数据classification 将accuracy 提高到100% |  |
| 各种数据classification 将accuracy 提高到100% |  |
| 各种数据classification 将accuracy 提高到100% |  |
| 各种数据classification 将accuracy 提高到100% |  |
| 各种数据classification 将accuracy 提高到100% |  |
| 各种数据classification 将accuracy 提高到100% |  |
| 各种数据classification 将accuracy 提高到100% |  |
| 各种数据classification 将accuracy 提高到100% |  |
| 各种数据classification 将accuracy 提高到100% |  |
| 各种数据classification 将accuracy 提高到100% |  |
| batch normalization的位置 |  |
| BP algorithms |  |
| Fully Connection |  |
| Forward |  |
| Forward |  |
| Back Propagation |  |
| 最简单版cnn |  |
| TODO LIST | |
| 自适应控制 | |
| Anchor-free |  |
| Anchor-free |  |
| anchor-basedanchor-free区别 |  |
| Anchor-free的局限 |  |
| anchor-free 的其他套路 |  |
| anchor-free 与anchor-based结合 |  |
| anchor-free 与anchor-based结合 |  |
| 为什么anchor-free能卷土重来 | |
| paddlepaddle |  |
| PAPER idea | |
| 自适应控制 | |
| su sudo |  |
 |  |
 | |
 |  |
 | |
| |
 |  |
 |  |
 |  |
 |  |
 |  |
 |  |
 |  |