22
33- [ ymir] ( https://github.com/IndustryEssentials/ymir )
44
5- - [ wiki ] ( https://github.com/modelai/ymir-executor-fork/wiki )
5+ - [ 说明文档 ] ( https://github.com/modelai/ymir-executor-fork/wiki )
66
7- ## ymir-1.1.0 官方镜像
7+ - [ ymir镜像 ] ( ./docs/official-docker-image.md )
88
9- - [ yolov4] ( https://github.com/modelai/ymir-executor-fork#det-yolov4-training )
10-
11- ```
12- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-yolov4-cu112-tmi
13-
14- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-yolov4-cu101-tmi
15- ```
16-
17- - [yolov5](https://github.com/modelai/ymir-executor-fork#det-yolov5-tmi)
18-
19- - [change log](./det-yolov5-tmi/README.md)
20-
21- ```
22- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-yolov5-cu111-tmi
23-
24- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-yolov5-cu102-tmi
25- ```
26-
27- - [mmdetection](https://github.com/modelai/ymir-executor-fork#det-mmdetection-tmi)
28-
29- - [change log](./det-mmdetection-tmi/README.md)
30-
31- ```
32- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-mmdet-cu111-tmi
33-
34- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-mmdet-cu102-tmi
35- ```
36-
37- - [detectron2](https://github.com/modelai/ymir-detectron2)
38-
39- - [change log](https://github.com/modelai/ymir-detectron2/blob/master/README.md)
40-
41- ```
42- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-detectron2-cu111-tmi
43- ```
44-
45- - [yolov7](https://github.com/modelai/ymir-yolov7)
46-
47- - [change log](https://github.com/modelai/ymir-yolov7/blob/main/ymir/README.md)
48-
49- ```
50- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-yolov7-cu111-tmi
51- ```
52-
53- - [vidt](https://github.com/modelai/ymir-vidt)
54-
55- - [change log](https://github.com/modelai/ymir-vidt/tree/main/ymir)
56-
57- ```
58- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-vidt-cu111-tmi
59- ```
60-
61- - [nanodet](https://github.com/modelai/ymir-nanodet/tree/ymir-dev)
62-
63- - [change log](https://github.com/modelai/ymir-nanodet/tree/ymir-dev/ymir)
64-
65- ```
66- docker pull youdaoyzbx/ymir-executor:ymir1.1.0-nanodet-cu111-tmi
67- ```
9+ - [ ymir 挖掘算法] ( ./docs/mining-images-overview.md )
6810
6911## 比较
7012
7315| yolov4 | ? | ✔️ | ❌ | darknet + mxnet | ❌ | local |
7416| yolov5 | ✔️ | ✔️ | ✔️ | pytorch | ✔️ | local+online |
7517| yolov7 | ✔️ | ✔️ | ✔️ | pytorch | ❌ | local+online |
76- | mmdetection | ✔️ | ✔️ | ✔️ | pytorch | ❌ | online |
18+ | mmdetection | ✔️ | ✔️ | ✔️ | pytorch | ❌ | local+ online |
7719| detectron2 | ✔️ | ✔️ | ✔️ | pytorch | ❌ | online |
7820| vidt | ? | ✔️ | ✔️ | pytorch | ❌ | online |
79- | nanodet | ✔️ | ✔️ | ❌ | pytorch_lightning | ❌ | online |
21+ | nanodet | ✔️ | ✔️ | ❌ | pytorch_lightning | ❌ | local+ online |
8022
8123- ` online ` 预训练权重可能在训练时通过网络下载
8224
@@ -112,6 +54,8 @@ gpu: single GeForce GTX 1080 Ti
11254
11355---
11456
57+ # 手动构建ymir镜像
58+
11559## det-yolov4-tmi
11660
11761- yolov4的训练、挖掘与推理镜像,采用mxnet与darknet框架
@@ -145,7 +89,7 @@ docker build -t ymir-executor/mmdet:cu111-tmi -f docker/Dockerfile.cuda111 .
14589
14690## live-code-executor
14791
148- - 可以通过`git_url`, `commit id` 或 `tag` 从网上clone代码到镜像并运行, 不推荐使用`branch`, 因为这样拉取的代码可能随时间变化, 实验结果不具备可重复性 .
92+ - 可以通过` git_url ` , ` commit id ` 或 ` tag ` 从网上clone代码到镜像并运行, 不推荐使用` branch ` , 因为这样拉取的代码可能随时间变化, 过程不具备可重复性 .
14993
15094- 参考 [ live-code] ( https://github.com/IndustryEssentials/ymir-remote-git )
15195
@@ -159,10 +103,14 @@ docker build -t ymir-executor/live-code:mxnet-tmi -f mxnet.dockerfile
159103
160104## 如何制作自己的ymir-executor
161105
106+ - [ 示例 ymir-executor] ( det-demo-tmi/README.md ) 从零到一,搭建自己的 ymir-executor
107+
162108- [ ymir-executor 制作指南] ( https://github.com/IndustryEssentials/ymir/blob/dev/dev_docs/ymir-dataset-zh-CN.md )
163109
164110- [ ymir-executor-sdk] ( https://github.com/modelai/ymir-executor-sdk ) ymir镜像开发辅助库
165111
112+ - [ ymir-executor-verifer] ( https://github.com/modelai/ymir-executor-verifier ) 调试与检测 ymir-executor
113+
166114## 如何导入预训练模型
167115
168116- [ 如何导入并精调外部模型] ( https://github.com/modelai/ymir-executor-fork/wiki/import-and-finetune-model )
@@ -189,7 +137,7 @@ docker build -t ymir-executor/live-code:mxnet-tmi -f mxnet.dockerfile
189137
190138## 关于cuda版本
191139
192- - 推荐主机安装11 .2以上的cuda版本, 使用11.1及以上的镜像
140+ - 推荐主机安装高版本驱动,支持11 .2以上的cuda版本, 使用11.1及以上的镜像
193141
194142- GTX3080/GTX3090不支持11.1以下的cuda,只能使用cuda11.1及以上的镜像
195143
0 commit comments