1. 图象分类#
数据集：ImageNet1000类
| Model | 压缩方法 | Top-1/Top-5 | 模型大小（MB） | FLOPs | 下载 |
|---|---|---|---|---|---|
| MobileNetV1 | - | 70.99%/89.68% | xx | xx | 下载链接 |
| MobileNetV1 | quant_psot | xx%/xx% | xx | - | 下载链接 |
| MobileNetV1 | quant_aware | xx%/xx% | xx | - | 下载链接 |
| MobileNetV1 | uniform -xx% | xx%/xx% | xx | xx | 下载链接 |
| MobileNetV1 | sensitive -xx% | xx%/xx% | xx | xx | 下载链接 |
| MobileNetV1 | ResNet50_vd1 | xx%/xx% | xx | - | 下载链接 |
| MobileNetV2 | - | 72.15%/90.65% | xx | xx | 下载链接 |
| MobileNetV2 | quant_post | xx%/xx% | xx | - | 下载链接 |
| MobileNetV2 | quant_aware | xx%/xx% | xx | - | 下载链接 |
| MobileNetV2 | uniform -xx% | xx%/xx% | xx | xx | 下载链接 |
| MobileNetV2 | sensitive -xx% | xx%/xx% | xx | xx | 下载链接 |
| MobileNetV2 | ResNet50_vd1 | xx%/xx% | xx | - | 下载链接 |
| ResNet34 | - | 74.57%/92.14% | xx | xx | 下载链接 |
| ResNet34 | uniform -xx% | xx%/xx% | xx | xx | 下载链接 |
| ResNet34 | auto -xx% | xx%/xx% | xx | xx | 下载链接 |
| ResNet50 | - | 76.50%/93.00% | xx | - | 下载链接 |
| ResNet50 | quant_post | xx%/xx% | xx | - | 下载链接 |
| ResNet50 | quant_aware | xx%/xx% | xx | - | 下载链接 |
| ResNet50 | ResNet1012 | xx%/xx% | xx | - | 下载链接 |
2. 目标检测#
数据集：COCO 2017
| Model | 压缩方法 | 数据集 | Image/GPU | 输入608 Box AP | 输入416 Box AP | 输入320 Box AP | 模型大小(MB) | FLOPs | 下载 |
|---|---|---|---|---|---|---|---|---|---|
| MobileNet-V1-YOLOv3 | - | Pasacl VOC | 8 | 76.2 | 76.7 | 75.3 | xx | - | 下载链接 |
| MobileNet-V1-YOLOv3 | uniform -xx% | Pasacl VOC | 8 | xx | xx | xx | xx | - | 下载链接 |
| MobileNet-V1-YOLOv3 | ResNet34-YOLOv3-VOC3 | Pasacl VOC | 16 | xx | xx | xx | xx | - | 下载链接 |
| MobileNet-V1-YOLOv3 | - | COCO | 8 | 29.3 | 29.3 | 27.1 | xx | xx | 下载链接 |
| MobileNet-V1-YOLOv3 | quant_post | COCO | 8 | xx | xx | xx | xx | - | 下载链接 |
| MobileNet-V1-YOLOv3 | quant_aware | COCO | 8 | xx | xx | xx | xx | - | 下载链接 |
| MobileNet-V1-YOLOv3 | uniform -xx% | COCO | 8 | xx | xx | xx | xx | xx | 下载链接 |
| MobileNet-V1-YOLOv3 | ResNet34-YOLOv3-COCO4 | COCO | 16 | xx | xx | xx | xx | - | 下载链接 |
| R50-dcn-YOLOv3 FP32 | - | COCO | 8 | 41.4 | xx | xx | xx | xx | 下载链接 |
| R50-dcn-YOLOv3 | quant_post | COCO | 8 | xx | xx | xx | xx | - | 下载链接 |
| R50-dcn-YOLOv3 | quant_aware | COCO | 8 | xx | xx | xx | xx | - | 下载链接 |
| R50-dcn-YOLOv3 | uniform -xx% | COCO | 8 | xx | xx | xx | xx | xx | 下载链接 |
数据集：WIDER-FACE
| Model | 压缩方法 | Image/GPU | 输入尺寸 | Easy/Medium/Hard | 模型大小（MB） | 下载 |
|---|---|---|---|---|---|---|
| BlazeFace | - | 8 | 640 | 0.915/0.892/0.797 | xx | 下载链接 |
| BlazeFace | quant_post | 8 | 640 | xx/xx/xx | xx | 下载链接 |
| BlazeFace | quant_aware | 8 | 640 | xx/xx/xx | xx | 下载链接 |
| BlazeFace-Lite | - | 8 | 640 | 0.909/0.885/0.781 | xx | 下载链接 |
| BlazeFace-Lite | quant_post | 8 | 640 | xx/xx/xx | xx | 下载链接 |
| BlazeFace-Lite | quant_aware | 8 | 640 | xx/xx/xx | xx | 下载链接 |
| BlazeFace-NAS | - | 8 | 640 | 0.837/0.807/0.658 | xx | 下载链接 |
| BlazeFace-NAS | quant_post | 8 | 640 | xx/xx/xx | xx | 下载链接 |
| BlazeFace-NAS | quant_aware | 8 | 640 | xx/xx/xx | xx | 下载链接 |
3. 图像分割#
数据集：Cityscapes
| Model | 压缩方法 | mIoU | 模型大小（MB） | FLOPs | 下载 |
|---|---|---|---|---|---|
| DeepLabv3+/MobileNetv1 | - | 63.26 | xx | - | 下载链接 |
| DeepLabv3+/MobileNetv1 | quant_post | xx | xx | - | 下载链接 |
| DeepLabv3+/MobileNetv1 | quant_aware | xx | xx | - | 下载链接 |
| DeepLabv3+/MobileNetv2 | - | 69.81 | xx | xx | 下载链接 |
| DeepLabv3+/MobileNetv2 | quant_post | xx | xx | - | 下载链接 |
| DeepLabv3+/MobileNetv2 | quant_aware | xx | xx | - | 下载链接 |
| DeepLabv3+/MobileNetv2 | prune -xx% | xx | xx | xx | 下载链接 |
Note
[1]：ResNet50_vd预训练模型Top-1/Top-5准确率分别为79.12%/94.44%
带_vd后缀代表开启了Mixup训练，Mixup相关介绍参考mixup: Beyond Empirical Risk Minimization
[2]：ResNet101预训练模型Top-1/Top-5准确率分别为77.56%/93.64% [3]：ResNet34-YOLOv3-VOC预训练模型的Box AP为82.6
[4]：ResNet34-YOLOv3-COCO预训练模型的Box AP为36.2