MODEL_ZOO.md 7.4 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
# Model Zoo and Benchmark
## Environment

- Python 2.7.1
- PaddlePaddle 1.5
- CUDA 9.0
- CUDNN 7.4
- NCCL 2.1.2

## Common settings

- All models below except SSD were trained on `coco_2017_train`, and tested on `coco_2017_val`.
- Batch Normalization layers in backbones are replaced by Affine Channel layers.
- Unless otherwise noted, all ResNet backbones adopt the [ResNet-B](https://arxiv.org/pdf/1812.01187) variant..
- For RCNN and RetinaNet models, only horizontal flipping data augmentation was used in the training phase and no augmentations were used in the testing phase.

## Training Schedules

J
jerrywgz 已提交
19
- We adopt exactly the same training schedules as [Detectron](https://github.com/facebookresearch/Detectron/blob/master/MODEL_ZOO.md#training-schedules).
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
- 1x indicates the schedule starts at a LR of 0.02 and is decreased by a factor of 10 after 60k and 80k iterations and eventually terminates at 90k iterations for minibatch size 16. For batch size 8, LR is decreased to 0.01, total training iterations are doubled, and the decay milestones are scaled by 2.
- 2x schedule is twice as long as 1x, with the LR milestones scaled accordingly.

## ImageNet Pretrained Models

The backbone models pretrained on ImageNet are available. All backbone models are pretrained on standard ImageNet-1k dataset and can be downloaded [here](https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification#supported-models-and-performances).

- Notes:  The ResNet50 model was trained with cosine LR decay schedule and can be downloaded here [here](https://paddle-imagenet-models-name.bj.bcebos.com/ResNet50_cos_pretrained.tar).

## Baselines

### Faster & Mask R-CNN

| Backbone             | Type           | Img/gpu | Lr schd | Box AP | Mask AP |                           Download                           |
| :------------------- | :------------- | :-----: | :-----: | :----: | :-----: | :----------------------------------------------------------: |
J
jerrywgz 已提交
35 36 37
| ResNet50             | Faster         |    1    |   1x    |  35.2  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_1x.tar) |
| ResNet50             | Faster         |    1    |   2x    |  37.1  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_2x.tar) |
| ResNet50             | Mask           |    1    |   1x    |  36.5  |  32.2   | [model](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r50_1x.tar) |
38 39 40
| ResNet50             | Mask           |    1    |   2x    |        |         |                          [model]()                           |
| ResNet50-D           | Faster         |    1    |   1x    |  36.4  |    -    | [model](ttps://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_vd_1x.tar) |
| ResNet50-FPN         | Faster         |    2    |   1x    |  37.2  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_fpn_1x.tar) |
J
jerrywgz 已提交
41
| ResNet50-FPN         | Faster         |    2    |   2x    |  37.7  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_fpn_2x.tar) |
42
| ResNet50-FPN         | Mask           |    2    |   1x    |  37.9  |  34.2   | [model](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r50_fpn_1x.tar) |
J
jerrywgz 已提交
43 44
| ResNet50-FPN         | Cascade Faster |    2    |   1x    |  40.9  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/cascade_rcnn_r50_fpn_1x.tar) |
| ResNet50-D-FPN       | Faster         |    2    |   2x    |  38.9  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r50_vd_fpn_2x.tar) |
45
| ResNet50-D-FPN       | Mask           |    2    |   2x    |  39.8  |  35.4   | [model](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r50_vd_fpn_2x.tar) |
J
jerrywgz 已提交
46 47 48 49 50 51 52
| ResNet101            | Faster         |    1    |   1x    |  38.3  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r101_1x.tar) |
| ResNet101-FPN        | Faster         |    1    |   1x    |  38.7  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r101_fpn_1x.tar) |
| ResNet101-FPN        | Faster         |    1    |   2x    |  39.1  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r101_fpn_2x.tar) |
| ResNet101-FPN        | Mask           |    1    |   1x    |  39.5  |  35.2   | [model](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_r101_fpn_1x.tar) |
| ResNet101-D-FPN      | Faster         |    1    |   1x    |  40.0  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r101_fpn_1x.tar) |
| ResNet101-D-FPN      | Faster         |    1    |   2x    |  40.6  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_r101_fpn_2x.tar) |
| SENet154-D-FPN       | Faster         |    1    |  1.44x  |  43.5  |    -    | [model](https://paddlemodels.bj.bcebos.com/object_detection/faster_rcnn_se154_fpn_s1x.tar) |
53 54 55 56
| SENet154-D-FPN       | Mask           |    1    |  1.44x  |  44.0  |  38.7   | [model](https://paddlemodels.bj.bcebos.com/object_detection/mask_rcnn_se154_vd_fpn_s1x.tar) |

### Yolo v3

K
Kaipeng Deng 已提交
57 58 59 60 61 62 63 64 65 66 67 68 69 70
| Backbone     | Size | Img/gpu | Lr schd | Box AP | Download  |
| :----------- | :--: | :-----: | :-----: | :----: | :-------: |
| DarkNet53    | 608  |    8    |   120e  |  38.9  | [model](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_darknet.tar) |
| DarkNet53    | 416  |    8    |   120e  |  37.5  | [model](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_darknet.tar) |
| DarkNet53    | 320  |    8    |   120e  |  34.8  | [model](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_darknet.tar) |
| MobileNet-V1 | 608  |    8    |   120e  |  29.3  | [model](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v1.tar) |
| MobileNet-V1 | 416  |    8    |   120e  |  29.3  | [model](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v1.tar) |
| MobileNet-V1 | 320  |    8    |   120e  |  27.1  | [model](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v1.tar) |
| ResNet34     | 608  |    8    |   120e  |  36.2  | [model](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_r34.tar) |
| ResNet34     | 416  |    8    |   120e  |  34.3  | [model](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_r34.tar) |
| ResNet34     | 320  |    8    |   120e  |  31.4  | [model](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_r34.tar) |

**NOTE**: Yolo v3 trained in 8 GPU with total batch size as 64. Yolo v3 training data augmentations: mixup image, 
random distort image, random crop image, random expand image, random interpolate, random flip image.
71 72 73 74 75

### RetinaNet

| Backbone     | Size | Lr schd | Box AP | Download  |
| :----------- | :--: | :-----: | :----: | :-------: |
J
jerrywgz 已提交
76 77
| ResNet50-FPN | 300  |  120e   |  36.0  | [model](https://paddlemodels.bj.bcebos.com/object_detection/retinanet_r50_fpn_1x.tar) |
| ResNet101-FPN | 300  |  120e   |  37.3  | [model](https://paddlemodels.bj.bcebos.com/object_detection/retinanet_r101_fpn_1x.tar) |
78

J
jerrywgz 已提交
79
### SSD on PascalVOC
80

K
Kaipeng Deng 已提交
81 82 83 84 85 86
| Backbone     | Size | Img/gpu | Lr schd | Box AP | Download  |
| :----------- | :--: | :-----: | :-----: | :----: | :-------: |
| MobileNet v1 | 300  |    32   |   120e  |  73.2  | [model](https://paddlemodels.bj.bcebos.com/object_detection/ssd_mobilenet_v1_voc.tar) |

**NOTE**: SSD trained in 2 GPU with totoal batch size as 64. SSD training data augmentations: random distort image, 
random crop image, random expand image, random flip image.