Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

perf: 在深眠中使用目标检测 #204

Merged
merged 2 commits into from
Oct 1, 2024
Merged

perf: 在深眠中使用目标检测 #204

merged 2 commits into from
Oct 1, 2024

Conversation

lamb203
Copy link
Contributor

@lamb203 lamb203 commented Oct 1, 2024

看最近深眠好像老有问题,随便练了个模型(也有可能是底模效果太好?),上次和这次的深眠都挂过去了,要效果不行就回退吧ᵕ᷄≀ ̠ᵕ᷅

@MistEO
Copy link
Member

MistEO commented Oct 1, 2024

这是哪个底模,用不着这么大的参数量吧

@lamb203
Copy link
Contributor Author

lamb203 commented Oct 1, 2024

就yolov8n最小的那个

@MistEO
Copy link
Member

MistEO commented Oct 1, 2024

yolov8 最小的都 11M 了啊?彳亍吧

@lamb203
Copy link
Contributor Author

lamb203 commented Oct 1, 2024

第一次练,数据集也没多大呢

"recognition": "NeuralNetworkDetect",
"model": "LimbDetection.onnx",
"expected": 0,
"threshold": 0.1,
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

MaaXYZ/MaaFramework#367

这个阈值现在是有 bug 的,实际不 work

可能在 CI 中找个新版本试一下 0.1 能不能行

Copy link
Contributor Author

@lamb203 lamb203 Oct 1, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

1999的深眠背景还是挺干净的,一个页面只有两个要点,我是用maafw2.0.0测试的

Copy link
Member

@MistEO MistEO Oct 1, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

我的意思是,现在的 MaaFW 有 bug,你设阈值压根不生效,实际跑的还是 0.3。等新版本了能生效了,0.1 太低了可能有问题。

要么你就把这行删了就用默认值(0.3),要么你去 CI 里下个新版本试一下 0.1 能不能用

@lamb203
Copy link
Contributor Author

lamb203 commented Oct 1, 2024

原本只有6M,换成onnx导出就变成11M了

@MistEO MistEO merged commit 6148ac8 into MaaXYZ:main Oct 1, 2024
11 checks passed
@MistEO MistEO changed the title feat: 在深眠中使用目标检测 perf: 在深眠中使用目标检测 Oct 1, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants