mirror of https://github.com/open-mmlab/mmyolo.git
90 lines
3.4 KiB
Markdown
90 lines
3.4 KiB
Markdown
|
# 使用 mim 跨库调用其他 OpenMMLab 仓库的脚本
|
|||
|
|
|
|||
|
|
```{note}
|
|||
|
|
1. 目前暂不支持跨库调用所有脚本，正在修复中。等修复完成，本文档会添加更多的例子。
|
|||
|
|
2. 绘制 mAP 和 计算平均训练速度 两项功能在 MMDetection dev-3.x 分支中修复，目前需要通过源码安装该分支才能成功调用。
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
## 日志分析
|
|||
|
|
|
|||
|
|
### 曲线图绘制
|
|||
|
|
|
|||
|
|
MMDetection 中的 `tools/analysis_tools/analyze_logs.py` 可利用指定的训练 log 文件绘制 loss/mAP 曲线图， 第一次运行前请先运行 `pip install seaborn` 安装必要依赖。
|
|||
|
|
|
|||
|
|
```shell
|
|||
|
|
mim run mmdet analyze_logs plot_curve \
|
|||
|
|
${LOG} \ # 日志文件路径
|
|||
|
|
[--keys ${KEYS}] \ # 需要绘制的指标，默认为 'bbox_mAP'
|
|||
|
|
[--start-epoch ${START_EPOCH}] # 起始的 epoch，默认为 1
|
|||
|
|
[--eval-interval ${EVALUATION_INTERVAL}] \ # 评估间隔，默认为 1
|
|||
|
|
[--title ${TITLE}] \ # 图片标题，无默认值
|
|||
|
|
[--legend ${LEGEND}] \ # 图例，默认为 None
|
|||
|
|
[--backend ${BACKEND}] \ # 绘制后端，默认为 None
|
|||
|
|
[--style ${STYLE}] \ # 绘制风格，默认为 'dark'
|
|||
|
|
[--out ${OUT_FILE}] # 输出文件路径
|
|||
|
|
# [] 代表可选参数，实际输入命令行时，不用输入 []
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
样例：
|
|||
|
|
|
|||
|
|
- 绘制分类损失曲线图
|
|||
|
|
|
|||
|
|
```shell
|
|||
|
|
mim run mmdet analyze_logs plot_curve \
|
|||
|
|
yolov5_s-v61_syncbn_fast_8xb16-300e_coco_20220918_084700.log.json \
|
|||
|
|
--keys loss_cls \
|
|||
|
|
--legend loss_cls
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
<img src="https://user-images.githubusercontent.com/27466624/204747359-754555df-1f97-4d5c-87ca-9ad3a0badcce.png" width="600"/>
|
|||
|
|
|
|||
|
|
- 绘制分类损失、回归损失曲线图，保存图片为对应的 pdf 文件
|
|||
|
|
|
|||
|
|
```shell
|
|||
|
|
mim run mmdet analyze_logs plot_curve \
|
|||
|
|
yolov5_s-v61_syncbn_fast_8xb16-300e_coco_20220918_084700.log.json \
|
|||
|
|
--keys loss_cls loss_bbox \
|
|||
|
|
--legend loss_cls loss_bbox \
|
|||
|
|
--out losses_yolov5_s.pdf
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
<img src="https://user-images.githubusercontent.com/27466624/204748560-2d17ce4b-fb5f-4732-a962-329109e73aad.png" width="600"/>
|
|||
|
|
|
|||
|
|
- 在同一图像中比较两次运行结果的 bbox mAP
|
|||
|
|
|
|||
|
|
```shell
|
|||
|
|
mim run mmdet analyze_logs plot_curve \
|
|||
|
|
yolov5_s-v61_syncbn_fast_8xb16-300e_coco_20220918_084700.log.json \
|
|||
|
|
yolov5_n-v61_syncbn_fast_8xb16-300e_coco_20220919_090739.log.json \
|
|||
|
|
--keys bbox_mAP \
|
|||
|
|
--legend yolov5_s yolov5_n \
|
|||
|
|
--eval-interval 10 # 注意评估间隔必须和训练时设置的一致，否则会报错
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
<img src="https://user-images.githubusercontent.com/27466624/204748704-21db9f9e-386e-449c-91c7-2ce3f8b51f24.png" width="600"/>
|
|||
|
|
|
|||
|
|
### 计算平均训练速度
|
|||
|
|
|
|||
|
|
```shell
|
|||
|
|
mim run mmdet analyze_logs cal_train_time \
|
|||
|
|
${LOG} \ # 日志文件路径
|
|||
|
|
[--include-outliers] # 计算时包含每个 epoch 的第一个数据
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
样例：
|
|||
|
|
|
|||
|
|
```shell
|
|||
|
|
mim run mmdet analyze_logs cal_train_time \
|
|||
|
|
yolov5_s-v61_syncbn_fast_8xb16-300e_coco_20220918_084700.log.json
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
输出以如下形式展示：
|
|||
|
|
|
|||
|
|
```text
|
|||
|
|
-----Analyze train time of yolov5_s-v61_syncbn_fast_8xb16-300e_coco_20220918_084700.log.json-----
|
|||
|
|
slowest epoch 278, average time is 0.1705 s/iter
|
|||
|
|
fastest epoch 300, average time is 0.1510 s/iter
|
|||
|
|
time std over epochs is 0.0026
|
|||
|
|
average iter time: 0.1556 s/iter
|
|||
|
|
```
|