You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This matches the description in your paper. However, I see most other works of rectified flow do not use this sigma_t and they simply use the mean value mu_t as the sampled x_t. I wonder if you have explored on how big influence this sigma_t has on the model performance, and what is the appropriate range of values for sigma_t ? Thanks for your help in advance.
The text was updated successfully, but these errors were encountered:
Thank you for the question, and I think this is a really good one. Honestly I don't have a solid answer to this, because in the conditional flow matching paper (https://arxiv.org/abs/2302.00482v1) they have the small $\sigma$, but in rectified flow (https://arxiv.org/abs/2209.03003) this $\sigma$ does not exist. From a mathematical perspective, the boundary condition of the conditional probability field $p(x_t|x_1,x_0)$ requires the $\sigma$ to be small, and it does not seem to have a great theoretical impact if we just set it to 0.
Personally, I have not done investigations on this tricky $\sigma$, but some time ago someone told me setting $\sigma$ to a smaller value than the current one "seemed to lead to worse performance". If this observation is true, then I guess a non-zero $\sigma$ helps to "smooth" the flow matching trajectory so that the model does not only learn on the line between $x_1$ and $x_0$, but also the regions nearby. This is just an intuition though, and more empirical evidence is needed to verify this.
Hi, thanks for your great work. I notice that you add a small gaussian noise when sampling
x_t
inCFM
:This matches the description in your paper. However, I see most other works of rectified flow do not use this
sigma_t
and they simply use the mean valuemu_t
as the sampledx_t
. I wonder if you have explored on how big influence thissigma_t
has on the model performance, and what is the appropriate range of values forsigma_t
? Thanks for your help in advance.The text was updated successfully, but these errors were encountered: