drhead

Adding what I commented on the duplicate issue referenced above, for anyone who wants to attempt this: Support could be added by anyone willing to create an SKSE/F4SE (etc) plugin....

> > > From where are you sourcing generic diversity minimum range requirements? The range requirements are listed in the codex.

> > > We don't receive journal information about the biological signals on a planet until after it has been mapped. Likewise, 3rd party databases like EDSM do not publish...

Here's the journal file: https://drive.google.com/file/d/1R8b3Po8GJwNANYUw9vCyKUWN6Za8cXj4/view?usp=sharing

> Thanks for implementing this to A1111. Could I ask for your preferred settings for SD 1.5 non-ZSNR models, especially for Schedule Sigma Max and the Negative Guidance minimum sigma?...

I will fix its implementation on the other code paths. Here are a few grids for comparing the zero SNR noise schedule, generated using ZeroDiffusion 0.9. The first two are...

With further testing I am getting this error: `/home/drhead/miniconda3/envs/automatic/lib/python3.10/site-packages/torchsde/_brownian/brownian_interval.py:601: UserWarning: Should have ta>=t0 but got ta=0.0291675366461277 and t0=0.029168. warnings.warn(f"Should have ta>=t0 but got ta={ta} and t0={self._start}.")` Give me some time...

Current implementation also seems to break Heun, but not Restart for some odd reason. Need to investigate.

I have identified the root cause for some samplers breaking. It appears that Heun, DDIM, and UniPC use other values that can be derived from `alphas_cumprod` at various points in...

I went forward with re-deriving the values in the one place that used those values and now the DDIM sampler works more or less as expected. I have further investigated...