We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
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
你好,感谢你出色的工作!我在使用你的项目进行自定义数据的训练时(only custom-depth.txt),遇到一些问题想请教。训练损失看起来正常,但是在前几个epoch验证的时候,我发现输出的深度的min和max差距很小,像下图这样: 请问这种情况你初始化训练遇到过吗,还是说我需要针对自定义数据集进行一些参数的调整?
The text was updated successfully, but these errors were encountered:
我们确实遇到过这种情况,可能是因为初始化的时候深度预测值偏小(像图中都小于1),导致将参考深度图投影到其他图片时超出了范围。可以试着对参数进行一些调整,比如最近采样距离。
Sorry, something went wrong.
十分感谢你回复!请问具体这块是在哪儿进行调整呢,在sample_ray的函数内部进行修改吗?
在configs里面调整min_depth
我也遇到了同样的问题, 我调整多个min_depth(从0.0001~2), 训练过程中预测的depth一直处于较小的范围, 很难训出模型. 请问一下二位怎么解决的? @superjuices @LinShan-Bin
No branches or pull requests
你好,感谢你出色的工作!我在使用你的项目进行自定义数据的训练时(only custom-depth.txt),遇到一些问题想请教。训练损失看起来正常,但是在前几个epoch验证的时候,我发现输出的深度的min和max差距很小,像下图这样:

请问这种情况你初始化训练遇到过吗,还是说我需要针对自定义数据集进行一些参数的调整?
The text was updated successfully, but these errors were encountered: