每一个模型下 config.py为超参数 运行EMSS下的preprocess.py生成处理的pkl文件 之后运行即可
结果如下
| model | CTR (%) | CVR (%) | AVG (%) |
|---|---|---|---|
| MTL | 67.34 | 59.81 | 63.57 |
| EMSS | 68.68 | 61.49 | 65.08 |
| MMOE | 68.71 | 60.17 | 64.44 |
| PLE | 68.93 | 60.29 | 64.61 |
| M3oE | 68.76 | 61.23 | 65.13 |
silliye/simpleM3oE
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|