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
作者您好!! 文中提出的损失函数,是将实际预测的Xt-1和最大似然估计推导出的理想最优Xt-1做损失。
我在想可不可以用ground truth噪声采样得到的Xt-1来做损失,即用generate_random_states函数中的实际noises算出的score,传入reverse_sde_step函数中得到的Xt-1,来和网络预测的噪声采样得到的Xt-1做损失,这样合理吗?
用generate_random_states函数中的noises算出的score,传入reverse_sde_step函数中得到的Xt-1,和文中根据最大似然估计推导出的最优Xt-1有什么区别,这个问题想请教下您的看法。
The text was updated successfully, but these errors were encountered:
你好!我大概想了一下,利用score得到的Xt_1是forward过程中的path,这可能不同于maximum likelihood得到的path。这个想法比较有意思不过我也没有做过实验,你可以试一下 ^_^.
Sorry, something went wrong.
好的,谢谢!
No branches or pull requests
作者您好!!
文中提出的损失函数,是将实际预测的Xt-1和最大似然估计推导出的理想最优Xt-1做损失。
我在想可不可以用ground truth噪声采样得到的Xt-1来做损失,即用generate_random_states函数中的实际noises算出的score,传入reverse_sde_step函数中得到的Xt-1,来和网络预测的噪声采样得到的Xt-1做损失,这样合理吗?
用generate_random_states函数中的noises算出的score,传入reverse_sde_step函数中得到的Xt-1,和文中根据最大似然估计推导出的最优Xt-1有什么区别,这个问题想请教下您的看法。
The text was updated successfully, but these errors were encountered: