更换文档检测模型
This commit is contained in:
176
paddle_detection/configs/face_detection/README.md
Normal file
176
paddle_detection/configs/face_detection/README.md
Normal file
@@ -0,0 +1,176 @@
|
||||
# 人脸检测模型
|
||||
|
||||
## 简介
|
||||
`face_detection`中提供高效、高速的人脸检测解决方案，包括最先进的模型和经典模型。
|
||||
|
||||
|
||||
|
||||
## 模型库
|
||||
|
||||
#### WIDER-FACE数据集上的mAP
|
||||
|
||||
| 网络结构 | 输入尺寸 | 图片个数/GPU | 学习率策略 | Easy/Medium/Hard Set | 预测时延（SD855）| 模型大小(MB) | 下载 | 配置文件 |
|
||||
|:------------:|:--------:|:----:|:-------:|:-------:|:---------:|:----------:|:---------:|:--------:|
|
||||
| BlazeFace | 640 | 8 | 1000e | 0.885 / 0.855 / 0.731 | - | 0.472 |[下载链接](https://paddledet.bj.bcebos.com/models/blazeface_1000e.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/face_detection/blazeface_1000e.yml) |
|
||||
| BlazeFace-FPN-SSH | 640 | 8 | 1000e | 0.907 / 0.883 / 0.793 | - | 0.479 |[下载链接](https://paddledet.bj.bcebos.com/models/blazeface_fpn_ssh_1000e.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/face_detection/blazeface_fpn_ssh_1000e.yml) |
|
||||
|
||||
**注意:**
|
||||
- 我们使用多尺度评估策略得到`Easy/Medium/Hard Set`里的mAP。具体细节请参考[在WIDER-FACE数据集上评估](#在WIDER-FACE数据集上评估)。
|
||||
|
||||
## 快速开始
|
||||
|
||||
### 数据准备
|
||||
我们使用[WIDER-FACE数据集](http://shuoyang1213.me/WIDERFACE/)进行训练和模型测试，官方网站提供了详细的数据介绍。
|
||||
- WIDER-Face数据源:
|
||||
使用如下目录结构加载`wider_face`类型的数据集：
|
||||
|
||||
```
|
||||
dataset/wider_face/
|
||||
├── wider_face_split
|
||||
│ ├── wider_face_train_bbx_gt.txt
|
||||
│ ├── wider_face_val_bbx_gt.txt
|
||||
├── WIDER_train
|
||||
│ ├── images
|
||||
│ │ ├── 0--Parade
|
||||
│ │ │ ├── 0_Parade_marchingband_1_100.jpg
|
||||
│ │ │ ├── 0_Parade_marchingband_1_381.jpg
|
||||
│ │ │ │ ...
|
||||
│ │ ├── 10--People_Marching
|
||||
│ │ │ ...
|
||||
├── WIDER_val
|
||||
│ ├── images
|
||||
│ │ ├── 0--Parade
|
||||
│ │ │ ├── 0_Parade_marchingband_1_1004.jpg
|
||||
│ │ │ ├── 0_Parade_marchingband_1_1045.jpg
|
||||
│ │ │ │ ...
|
||||
│ │ ├── 10--People_Marching
|
||||
│ │ │ ...
|
||||
```
|
||||
|
||||
- 手动下载数据集：
|
||||
要下载WIDER-FACE数据集，请运行以下命令：
|
||||
```
|
||||
cd dataset/wider_face && ./download_wider_face.sh
|
||||
```
|
||||
|
||||
### 参数配置
|
||||
基础模型的配置可以参考`configs/face_detection/_base_/blazeface.yml`；
|
||||
改进模型增加FPN和SSH的neck结构，配置文件可以参考`configs/face_detection/_base_/blazeface_fpn.yml`，可以根据需求配置FPN和SSH，具体如下：
|
||||
```yaml
|
||||
BlazeNet:
|
||||
blaze_filters: [[24, 24], [24, 24], [24, 48, 2], [48, 48], [48, 48]]
|
||||
double_blaze_filters: [[48, 24, 96, 2], [96, 24, 96], [96, 24, 96],
|
||||
[96, 24, 96, 2], [96, 24, 96], [96, 24, 96]]
|
||||
act: hard_swish #配置backbone中BlazeBlock的激活函数，基础模型为relu，增加FPN和SSH时需使用hard_swish
|
||||
|
||||
BlazeNeck:
|
||||
neck_type : fpn_ssh #可选only_fpn、only_ssh和fpn_ssh
|
||||
in_channel: [96,96]
|
||||
```
|
||||
|
||||
|
||||
|
||||
### 训练与评估
|
||||
训练流程与评估流程方法与其他算法一致，请参考[GETTING_STARTED_cn.md](../../docs/tutorials/GETTING_STARTED_cn.md)。
|
||||
**注意:** 人脸检测模型目前不支持边训练边评估。
|
||||
|
||||
#### 在WIDER-FACE数据集上评估
|
||||
- 步骤一：评估并生成结果文件：
|
||||
```shell
|
||||
python -u tools/eval.py -c configs/face_detection/blazeface_1000e.yml \
|
||||
-o weights=output/blazeface_1000e/model_final \
|
||||
multi_scale=True
|
||||
```
|
||||
设置`multi_scale=True`进行多尺度评估，评估完成后，将在`output/pred`中生成txt格式的测试结果。
|
||||
|
||||
- 步骤二：下载官方评估脚本和Ground Truth文件：
|
||||
```
|
||||
wget http://mmlab.ie.cuhk.edu.hk/projects/WIDERFace/support/eval_script/eval_tools.zip
|
||||
unzip eval_tools.zip && rm -f eval_tools.zip
|
||||
```
|
||||
|
||||
- 步骤三：开始评估
|
||||
|
||||
方法一：python评估：
|
||||
```
|
||||
git clone https://github.com/wondervictor/WiderFace-Evaluation.git
|
||||
cd WiderFace-Evaluation
|
||||
# 编译
|
||||
python3 setup.py build_ext --inplace
|
||||
# 开始评估
|
||||
python3 evaluation.py -p /path/to/PaddleDetection/output/pred -g /path/to/eval_tools/ground_truth
|
||||
```
|
||||
|
||||
方法二：MatLab评估：
|
||||
```
|
||||
# 在`eval_tools/wider_eval.m`中修改保存结果路径和绘制曲线的名称：
|
||||
pred_dir = './pred';
|
||||
legend_name = 'Paddle-BlazeFace';
|
||||
|
||||
`wider_eval.m` 是评估模块的主要执行程序。运行命令如下：
|
||||
matlab -nodesktop -nosplash -nojvm -r "run wider_eval.m;quit;"
|
||||
```
|
||||
|
||||
### Python脚本预测
|
||||
为了支持二次开发，这里提供通过Python脚本使用Paddle Detection whl包来进行预测的示例。
|
||||
```python
|
||||
import cv2
|
||||
import paddle
|
||||
import numpy as np
|
||||
from ppdet.core.workspace import load_config
|
||||
from ppdet.engine import Trainer
|
||||
from ppdet.metrics import get_infer_results
|
||||
from ppdet.data.transform.operators import NormalizeImage, Permute
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
# 准备基础的参数
|
||||
config_path = 'PaddleDetection/configs/face_detection/blazeface_1000e.yml'
|
||||
cfg = load_config(config_path)
|
||||
weight_path = 'PaddleDetection/output/blazeface_1000e.pdparams'
|
||||
infer_img_path = 'PaddleDetection/demo/hrnet_demo.jpg'
|
||||
cfg.weights = weight_path
|
||||
bbox_thre = 0.8
|
||||
paddle.set_device('gpu')
|
||||
# 创建所需的类
|
||||
trainer = Trainer(cfg, mode='test')
|
||||
trainer.load_weights(cfg.weights)
|
||||
trainer.model.eval()
|
||||
normaler = NormalizeImage(mean=[123, 117, 104], std=[127.502231, 127.502231, 127.502231], is_scale=False)
|
||||
permuter = Permute()
|
||||
# 进行图片读取
|
||||
im = cv2.imread(infer_img_path)
|
||||
im = cv2.cvtColor(im, cv2.COLOR_BGR2RGB)
|
||||
# 准备数据字典
|
||||
data_dict = {'image': im}
|
||||
data_dict = normaler(data_dict)
|
||||
data_dict = permuter(data_dict)
|
||||
h, w, c = im.shape
|
||||
data_dict['im_id'] = paddle.Tensor(np.array([[0]]))
|
||||
data_dict['im_shape'] = paddle.Tensor(np.array([[h, w]], dtype=np.float32))
|
||||
data_dict['scale_factor'] = paddle.Tensor(np.array([[1., 1.]], dtype=np.float32))
|
||||
data_dict['image'] = paddle.Tensor(data_dict['image'].reshape((1, c, h, w)))
|
||||
data_dict['curr_iter'] = paddle.Tensor(np.array([0]))
|
||||
# 进行预测
|
||||
outs = trainer.model(data_dict)
|
||||
# 对预测的数据进行后处理得到最终的bbox信息
|
||||
for key in ['im_shape', 'scale_factor', 'im_id']:
|
||||
outs[key] = data_dict[key]
|
||||
for key, value in outs.items():
|
||||
outs[key] = value.numpy()
|
||||
clsid2catid, catid2name = {0: 'face'}, {0: 0}
|
||||
batch_res = get_infer_results(outs, clsid2catid)
|
||||
bbox = [sub_dict for sub_dict in batch_res['bbox'] if sub_dict['score'] > bbox_thre]
|
||||
print(bbox)
|
||||
```
|
||||
|
||||
## Citations
|
||||
|
||||
```
|
||||
@article{bazarevsky2019blazeface,
|
||||
title={BlazeFace: Sub-millisecond Neural Face Detection on Mobile GPUs},
|
||||
author={Valentin Bazarevsky and Yury Kartynnik and Andrey Vakunov and Karthik Raveendran and Matthias Grundmann},
|
||||
year={2019},
|
||||
eprint={1907.05047},
|
||||
archivePrefix={arXiv},
|
||||
```
|
||||
176
paddle_detection/configs/face_detection/README_en.md
Normal file
176
paddle_detection/configs/face_detection/README_en.md
Normal file
@@ -0,0 +1,176 @@
|
||||
# Face Detection Model
|
||||
|
||||
## Introduction
|
||||
`face_detection` High efficiency, high speed face detection solutions, including the most advanced models and classic models.
|
||||
|
||||
|
||||
|
||||
## Model Library
|
||||
|
||||
#### A mAP on the WIDERFACE dataset
|
||||
|
||||
| Network structure | size | images/GPUs | Learning rate strategy | Easy/Medium/Hard Set | Prediction delay（SD855）| Model size(MB) | Download | Configuration File |
|
||||
|:------------:|:--------:|:----:|:-------:|:-------:|:---------:|:----------:|:---------:|:--------:|
|
||||
| BlazeFace | 640 | 8 | 1000e | 0.885 / 0.855 / 0.731 | - | 0.472 |[link](https://paddledet.bj.bcebos.com/models/blazeface_1000e.pdparams) | [Configuration File](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/face_detection/blazeface_1000e.yml) |
|
||||
| BlazeFace-FPN-SSH | 640 | 8 | 1000e | 0.907 / 0.883 / 0.793 | - | 0.479 |[link](https://paddledet.bj.bcebos.com/models/blazeface_fpn_ssh_1000e.pdparams) | [Configuration File](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/face_detection/blazeface_fpn_ssh_1000e.yml) |
|
||||
|
||||
**Attention:**
|
||||
- We use a multi-scale evaluation strategy to get the mAP in `Easy/Medium/Hard Set`. Please refer to the [evaluation on the WIDER FACE dataset](#Evaluated-on-the-WIDER-FACE-Dataset) for details.
|
||||
|
||||
## Quick Start
|
||||
|
||||
### Data preparation
|
||||
We use [WIDER-FACE dataset](http://shuoyang1213.me/WIDERFACE/) for training and model tests, the official web site provides detailed data is introduced.
|
||||
- WIDER-Face data source:
|
||||
- Load a dataset of type `wider_face` using the following directory structure:
|
||||
```
|
||||
dataset/wider_face/
|
||||
├── wider_face_split
|
||||
│ ├── wider_face_train_bbx_gt.txt
|
||||
│ ├── wider_face_val_bbx_gt.txt
|
||||
├── WIDER_train
|
||||
│ ├── images
|
||||
│ │ ├── 0--Parade
|
||||
│ │ │ ├── 0_Parade_marchingband_1_100.jpg
|
||||
│ │ │ ├── 0_Parade_marchingband_1_381.jpg
|
||||
│ │ │ │ ...
|
||||
│ │ ├── 10--People_Marching
|
||||
│ │ │ ...
|
||||
├── WIDER_val
|
||||
│ ├── images
|
||||
│ │ ├── 0--Parade
|
||||
│ │ │ ├── 0_Parade_marchingband_1_1004.jpg
|
||||
│ │ │ ├── 0_Parade_marchingband_1_1045.jpg
|
||||
│ │ │ │ ...
|
||||
│ │ ├── 10--People_Marching
|
||||
│ │ │ ...
|
||||
```
|
||||
|
||||
- Manually download the dataset:
|
||||
To download the WIDER-FACE dataset, run the following command:
|
||||
```
|
||||
cd dataset/wider_face && ./download_wider_face.sh
|
||||
```
|
||||
|
||||
### Parameter configuration
|
||||
The configuration of the base model can be referenced to `configs/face_detection/_base_/blazeface.yml`；
|
||||
Improved model to add FPN and SSH neck structure, configuration files can be referenced to `configs/face_detection/_base_/blazeface_fpn.yml`, You can configure FPN and SSH as required
|
||||
```yaml
|
||||
BlazeNet:
|
||||
blaze_filters: [[24, 24], [24, 24], [24, 48, 2], [48, 48], [48, 48]]
|
||||
double_blaze_filters: [[48, 24, 96, 2], [96, 24, 96], [96, 24, 96],
|
||||
[96, 24, 96, 2], [96, 24, 96], [96, 24, 96]]
|
||||
act: hard_swish #Configure Blaze Block activation function in Backbone. The basic model is Relu. hard_swish is needed to add FPN and SSH
|
||||
|
||||
BlazeNeck:
|
||||
neck_type : fpn_ssh #only_fpn, only_ssh and fpn_ssh
|
||||
in_channel: [96,96]
|
||||
```
|
||||
|
||||
|
||||
|
||||
### Training and Evaluation
|
||||
The training process and evaluation process methods are consistent with other algorithms, please refer to [GETTING_STARTED_cn.md](../../docs/tutorials/GETTING_STARTED_cn.md)。
|
||||
**Attention:** Face detection models currently do not support training and evaluation.
|
||||
|
||||
#### Evaluated on the WIDER-FACE Dataset
|
||||
- Step 1: Evaluate and generate a result file:
|
||||
```shell
|
||||
python -u tools/eval.py -c configs/face_detection/blazeface_1000e.yml \
|
||||
-o weights=output/blazeface_1000e/model_final \
|
||||
multi_scale=True
|
||||
```
|
||||
Set `multi_scale=True` for multi-scale evaluation. After evaluation, test results in TXT format will be generated in `output/pred`.
|
||||
|
||||
- Step 2: Download the official evaluation script and Ground Truth file:
|
||||
```
|
||||
wget http://mmlab.ie.cuhk.edu.hk/projects/WIDERFace/support/eval_script/eval_tools.zip
|
||||
unzip eval_tools.zip && rm -f eval_tools.zip
|
||||
```
|
||||
|
||||
- Step 3: Start the evaluation
|
||||
|
||||
Method 1: Python evaluation:
|
||||
```
|
||||
git clone https://github.com/wondervictor/WiderFace-Evaluation.git
|
||||
cd WiderFace-Evaluation
|
||||
# compile
|
||||
python3 setup.py build_ext --inplace
|
||||
# Begin to assess
|
||||
python3 evaluation.py -p /path/to/PaddleDetection/output/pred -g /path/to/eval_tools/ground_truth
|
||||
```
|
||||
|
||||
Method 2: MatLab evaluation:
|
||||
```
|
||||
# Change the name of save result path and draw curve in `eval_tools/wider_eval.m`:
|
||||
pred_dir = './pred';
|
||||
legend_name = 'Paddle-BlazeFace';
|
||||
|
||||
`wider_eval.m` is the main implementation of the evaluation module. Run the following command:
|
||||
matlab -nodesktop -nosplash -nojvm -r "run wider_eval.m;quit;"
|
||||
```
|
||||
|
||||
### Use by Python Code
|
||||
In order to support development, here is an example of using the Paddle Detection whl package to make predictions through Python code.
|
||||
```python
|
||||
import cv2
|
||||
import paddle
|
||||
import numpy as np
|
||||
from ppdet.core.workspace import load_config
|
||||
from ppdet.engine import Trainer
|
||||
from ppdet.metrics import get_infer_results
|
||||
from ppdet.data.transform.operators import NormalizeImage, Permute
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
# prepare for the parameters
|
||||
config_path = 'PaddleDetection/configs/face_detection/blazeface_1000e.yml'
|
||||
cfg = load_config(config_path)
|
||||
weight_path = 'PaddleDetection/output/blazeface_1000e.pdparams'
|
||||
infer_img_path = 'PaddleDetection/demo/hrnet_demo.jpg'
|
||||
cfg.weights = weight_path
|
||||
bbox_thre = 0.8
|
||||
paddle.set_device('gpu')
|
||||
# create the class object
|
||||
trainer = Trainer(cfg, mode='test')
|
||||
trainer.load_weights(cfg.weights)
|
||||
trainer.model.eval()
|
||||
normaler = NormalizeImage(mean=[123, 117, 104], std=[127.502231, 127.502231, 127.502231], is_scale=False)
|
||||
permuter = Permute()
|
||||
# read the image file
|
||||
im = cv2.imread(infer_img_path)
|
||||
im = cv2.cvtColor(im, cv2.COLOR_BGR2RGB)
|
||||
# prepare for the data dict
|
||||
data_dict = {'image': im}
|
||||
data_dict = normaler(data_dict)
|
||||
data_dict = permuter(data_dict)
|
||||
h, w, c = im.shape
|
||||
data_dict['im_id'] = paddle.Tensor(np.array([[0]]))
|
||||
data_dict['im_shape'] = paddle.Tensor(np.array([[h, w]], dtype=np.float32))
|
||||
data_dict['scale_factor'] = paddle.Tensor(np.array([[1., 1.]], dtype=np.float32))
|
||||
data_dict['image'] = paddle.Tensor(data_dict['image'].reshape((1, c, h, w)))
|
||||
data_dict['curr_iter'] = paddle.Tensor(np.array([0]))
|
||||
# do the prediction
|
||||
outs = trainer.model(data_dict)
|
||||
# to do the postprocess to get the final bbox info
|
||||
for key in ['im_shape', 'scale_factor', 'im_id']:
|
||||
outs[key] = data_dict[key]
|
||||
for key, value in outs.items():
|
||||
outs[key] = value.numpy()
|
||||
clsid2catid, catid2name = {0: 'face'}, {0: 0}
|
||||
batch_res = get_infer_results(outs, clsid2catid)
|
||||
bbox = [sub_dict for sub_dict in batch_res['bbox'] if sub_dict['score'] > bbox_thre]
|
||||
print(bbox)
|
||||
```
|
||||
|
||||
|
||||
## Citations
|
||||
|
||||
```
|
||||
@article{bazarevsky2019blazeface,
|
||||
title={BlazeFace: Sub-millisecond Neural Face Detection on Mobile GPUs},
|
||||
author={Valentin Bazarevsky and Yury Kartynnik and Andrey Vakunov and Karthik Raveendran and Matthias Grundmann},
|
||||
year={2019},
|
||||
eprint={1907.05047},
|
||||
archivePrefix={arXiv},
|
||||
```
|
||||
45
paddle_detection/configs/face_detection/_base_/blazeface.yml
Normal file
45
paddle_detection/configs/face_detection/_base_/blazeface.yml
Normal file
@@ -0,0 +1,45 @@
|
||||
architecture: BlazeFace
|
||||
|
||||
BlazeFace:
|
||||
backbone: BlazeNet
|
||||
neck: BlazeNeck
|
||||
blaze_head: FaceHead
|
||||
post_process: BBoxPostProcess
|
||||
|
||||
BlazeNet:
|
||||
blaze_filters: [[24, 24], [24, 24], [24, 48, 2], [48, 48], [48, 48]]
|
||||
double_blaze_filters: [[48, 24, 96, 2], [96, 24, 96], [96, 24, 96],
|
||||
[96, 24, 96, 2], [96, 24, 96], [96, 24, 96]]
|
||||
act: relu
|
||||
|
||||
BlazeNeck:
|
||||
neck_type : None
|
||||
in_channel: [96,96]
|
||||
|
||||
FaceHead:
|
||||
in_channels: [96,96]
|
||||
anchor_generator: AnchorGeneratorSSD
|
||||
loss: SSDLoss
|
||||
|
||||
SSDLoss:
|
||||
overlap_threshold: 0.35
|
||||
|
||||
AnchorGeneratorSSD:
|
||||
steps: [8., 16.]
|
||||
aspect_ratios: [[1.], [1.]]
|
||||
min_sizes: [[16.,24.], [32., 48., 64., 80., 96., 128.]]
|
||||
max_sizes: [[], []]
|
||||
offset: 0.5
|
||||
flip: False
|
||||
min_max_aspect_ratios_order: false
|
||||
|
||||
BBoxPostProcess:
|
||||
decode:
|
||||
name: SSDBox
|
||||
nms:
|
||||
name: MultiClassNMS
|
||||
keep_top_k: 750
|
||||
score_threshold: 0.01
|
||||
nms_threshold: 0.3
|
||||
nms_top_k: 5000
|
||||
nms_eta: 1.0
|
||||
@@ -0,0 +1,45 @@
|
||||
architecture: BlazeFace
|
||||
|
||||
BlazeFace:
|
||||
backbone: BlazeNet
|
||||
neck: BlazeNeck
|
||||
blaze_head: FaceHead
|
||||
post_process: BBoxPostProcess
|
||||
|
||||
BlazeNet:
|
||||
blaze_filters: [[24, 24], [24, 24], [24, 48, 2], [48, 48], [48, 48]]
|
||||
double_blaze_filters: [[48, 24, 96, 2], [96, 24, 96], [96, 24, 96],
|
||||
[96, 24, 96, 2], [96, 24, 96], [96, 24, 96]]
|
||||
act: hard_swish
|
||||
|
||||
BlazeNeck:
|
||||
neck_type : fpn_ssh
|
||||
in_channel: [96,96]
|
||||
|
||||
FaceHead:
|
||||
in_channels: [48, 48]
|
||||
anchor_generator: AnchorGeneratorSSD
|
||||
loss: SSDLoss
|
||||
|
||||
SSDLoss:
|
||||
overlap_threshold: 0.35
|
||||
|
||||
AnchorGeneratorSSD:
|
||||
steps: [8., 16.]
|
||||
aspect_ratios: [[1.], [1.]]
|
||||
min_sizes: [[16.,24.], [32., 48., 64., 80., 96., 128.]]
|
||||
max_sizes: [[], []]
|
||||
offset: 0.5
|
||||
flip: False
|
||||
min_max_aspect_ratios_order: false
|
||||
|
||||
BBoxPostProcess:
|
||||
decode:
|
||||
name: SSDBox
|
||||
nms:
|
||||
name: MultiClassNMS
|
||||
keep_top_k: 750
|
||||
score_threshold: 0.01
|
||||
nms_threshold: 0.3
|
||||
nms_top_k: 5000
|
||||
nms_eta: 1.0
|
||||
@@ -0,0 +1,44 @@
|
||||
worker_num: 2
|
||||
TrainReader:
|
||||
inputs_def:
|
||||
num_max_boxes: 90
|
||||
sample_transforms:
|
||||
- Decode: {}
|
||||
- RandomDistort: {brightness: [0.5, 1.125, 0.875], random_apply: False}
|
||||
- RandomExpand: {fill_value: [123.675, 116.28, 103.53]}
|
||||
- RandomFlip: {}
|
||||
- CropWithDataAchorSampling: {
|
||||
anchor_sampler: [[1, 10, 1.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.2, 0.0]],
|
||||
batch_sampler: [
|
||||
[1, 50, 1.0, 1.0, 1.0, 1.0, 0.0, 0.0, 1.0, 0.0],
|
||||
[1, 50, 0.3, 1.0, 1.0, 1.0, 0.0, 0.0, 1.0, 0.0],
|
||||
[1, 50, 0.3, 1.0, 1.0, 1.0, 0.0, 0.0, 1.0, 0.0],
|
||||
[1, 50, 0.3, 1.0, 1.0, 1.0, 0.0, 0.0, 1.0, 0.0],
|
||||
[1, 50, 0.3, 1.0, 1.0, 1.0, 0.0, 0.0, 1.0, 0.0],
|
||||
],
|
||||
target_size: 640}
|
||||
- Resize: {target_size: [640, 640], keep_ratio: False, interp: 1}
|
||||
- NormalizeBox: {}
|
||||
- PadBox: {num_max_boxes: 90}
|
||||
batch_transforms:
|
||||
- NormalizeImage: {mean: [123, 117, 104], std: [127.502231, 127.502231, 127.502231], is_scale: false}
|
||||
- Permute: {}
|
||||
batch_size: 8
|
||||
shuffle: true
|
||||
drop_last: true
|
||||
|
||||
|
||||
EvalReader:
|
||||
sample_transforms:
|
||||
- Decode: {}
|
||||
- NormalizeImage: {mean: [123, 117, 104], std: [127.502231, 127.502231, 127.502231], is_scale: false}
|
||||
- Permute: {}
|
||||
batch_size: 1
|
||||
|
||||
|
||||
TestReader:
|
||||
sample_transforms:
|
||||
- Decode: {}
|
||||
- NormalizeImage: {mean: [123, 117, 104], std: [127.502231, 127.502231, 127.502231], is_scale: false}
|
||||
- Permute: {}
|
||||
batch_size: 1
|
||||
@@ -0,0 +1,21 @@
|
||||
epoch: 1000
|
||||
|
||||
LearningRate:
|
||||
base_lr: 0.001
|
||||
schedulers:
|
||||
- !PiecewiseDecay
|
||||
gamma: 0.1
|
||||
milestones:
|
||||
- 333
|
||||
- 800
|
||||
- !LinearWarmup
|
||||
start_factor: 0.3333333333333333
|
||||
steps: 500
|
||||
|
||||
OptimizerBuilder:
|
||||
optimizer:
|
||||
momentum: 0.0
|
||||
type: RMSProp
|
||||
regularizer:
|
||||
factor: 0.0005
|
||||
type: L2
|
||||
@@ -0,0 +1,9 @@
|
||||
_BASE_: [
|
||||
'../datasets/wider_face.yml',
|
||||
'../runtime.yml',
|
||||
'_base_/optimizer_1000e.yml',
|
||||
'_base_/blazeface.yml',
|
||||
'_base_/face_reader.yml',
|
||||
]
|
||||
weights: output/blazeface_1000e/model_final
|
||||
multi_scale_eval: True
|
||||
@@ -0,0 +1,9 @@
|
||||
_BASE_: [
|
||||
'../datasets/wider_face.yml',
|
||||
'../runtime.yml',
|
||||
'_base_/optimizer_1000e.yml',
|
||||
'_base_/blazeface_fpn.yml',
|
||||
'_base_/face_reader.yml',
|
||||
]
|
||||
weights: output/blazeface_fpn_ssh_1000e/model_final
|
||||
multi_scale_eval: True
|
||||
Reference in New Issue
Block a user