Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Integrate deep learning models for image classification | Backbone learning/comparison/magic modification project
| Date | Stars |
|---|---|
| 2026-07-24 | 1951 |
| 2026-07-25 | 1951 |
| 2026-07-28 | 1951 |
| 2026-07-30 | 1951 |
| 2026-07-31 | 1952 |
| 2026-08-06 | 1952 |
Today
— stars today
This week
+1 stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.05%/day
Awesome backbones for image classification
===========================
<div align="center">
[](https://space.bilibili.com/46880349)
[](https://github.com/Fafa-DL/Awesome-Backbones)
[](https://github.com/Fafa-DL/Awesome-Backbones)
</div>
## 写在前面
- 若训练效果不佳，首先需要调整学习率和Batch size，这俩超参很大程度上影响收敛。其次，从关闭图像增强手段（尤其小数据集）开始，有的图像增强方法会污染数据，如
  
  如何去除增强？如[efficientnetv2-b0](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/models/efficientnetv2/efficientnetv2_b0.py)配置文件中train_pipeline可更改为如下
```yaml
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(
type='RandomResizedCrop',
size=192,
efficientnet_style=True,
interpolation='bicubic'),
dict(type='Normalize', **img_norm_cfg),
dict(type='ImageToTensor', keys=['img']),
dict(type='ToTensor', keys=['gt_label']),
dict(type='Collect', keys=['img', 'gt_label'])
]
```
  若你的数据集提前已经将shape更改为网络要求的尺寸，那么`Resize`操作也可以去除。
## 更新日志
**`2025.01.17`**
- 支持转ONNX[#136](https://github.com/Fafa-DL/Awesome-Backbones/pull/136) @PurpleSky-NS
- 类别激活图相关脚本[#114](https://github.com/Fafa-DL/Awesome-Backbones/pull/114) @jackyjinjing
**`2024.09.06`**
- 修复高频反馈的**评估时结果浮动大**的问题
**`2023.12.02`**
- 新增Issue中多人提及的输出**Train Acc**与**Val loss**
- `metrics_outputs.csv`保存每周期`train_loss, train_acc, train_precision, train_recall, train_f1-score, val_loss, val_acc, val_precision, val_recall, val_f1-score`方便各位绘图
- 终端由原先仅输出**Val**相关metrics升级为Train与Val都输出
**`2023.08.05`**
- 新增**TinyViT**(预训练权重不匹配)、**DeiT3**、**EdgeNeXt**、**RevVisionTransformer**
**`2023.03.07`**
- 新增**MobileViT**、**DaViT**、**RepLKNet**、**BEiT**、**EVA**、**MixMIM**、**EfficientNetV2**
## 测试环境
- Pytorch 1.7.1+
- Python 3.6+
## 资料
|数据集|视频教程|人工智能技术探讨群|
|---|---|---|
|[`花卉数据集` 提取码：0zat](https://pan.baidu.com/s/1137y4l-J3AgyCiC_cXqIqw)|[点我跳转](https://www.bilibili.com/video/BV1SY411P7Nd)|[1群：78174903](https://jq.qq.com/?_wv=1027&k=lY5KVICA)<br/>[3群：584723646](https://jq.qq.com/?_wv=1027&k=bakez5Yz)
## 快速开始
- 遵循[环境搭建](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/datas/docs/Environment_setting.md)完成配置
- 下载[MobileNetV3-Small](https://download.openmmlab.com/mmclassification/v0/mobilenet_v3/convert/mobilenet_v3_small-8427ecf0.pth)权重至**datas**下
- **Awesome-Backbones**文件夹下终端输入
```bash
python tools/single_test.py datas/cat-dog.png models/mobilenet/mobilenet_v3_small.py --classes-map datas/imageNet1kAnnotation.txt
```
## 教程
- [环境搭建](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/datas/docs/Environment_setting.md)
- [数据集准备](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/datas/docs/Data_preparing.md)
- [配置文件解释](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/datas/docs/Configs_description.md)
- [训练](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/datas/docs/How_to_train.md)
- [模型评估](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/datas/docs/How_to_eval.md)
- [计算Flops&Params](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/datas/docs/Calculate_Flops.md)
- [添加新的模型组件](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/datas/docs/Add_modules.md)
- [类别激活图可视化](https://github.com/Fafa-DL/Awesome-Backbones/blob/main/datas/docs/CAM_visualization.md)
- [学习率Excerpt of 35,880 characters
Read on GitHub
Fafa Bro
114
4
3
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:0b776573792fb878, topic:deep-learning, topic:pytorch
matched fp:0b776573792fb878, topic:transformer
matched fp:0b776573792fb878, topic:image-classification, desc:image classification, readme:image classification