训练ConvNext版本的YOLOv7和YOLOX #46
lucasjinreal
started this conversation in
Ideas
Replies: 2 comments
-
Beta Was this translation helpful? Give feedback.
0 replies
-
@PeterSmith1 可以pull一下。应该可以训练了。 第一步我们先测试从 0, 1, 2三层拿FPN feature,可能不一定work,先训练一波看看。后面试一下从 1, 2, 3 拿feature. 可以继续再这里同步训练中间结果,如果情况不对就再调整 |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
这里记录社区贡献者训练过程。
配置
主要是针对Convnext来做性能探索,由于Convnext同时兼备卷积的效率,以及transformer的性能,非常适合探索一波它能否超越我们常用的baseline resnet50,同时是否又能够在效率上超过卷积,需要深入试验探索。
该实验使用配置:https://github.com/jinfagang/yolov7/blob/main/configs/coco/yolox/yolox_convnext.yaml
训练任务启动脚本:
需要解决定位的问题:
Beta Was this translation helpful? Give feedback.
All reactions