Open JeffreyBull76 opened 11 months ago
All of the samplers that inject some kind of randomness somewhere essentially don't work, So that's all ancestral samplers and SDE samplers. I don't quite know what causes your other issues. This is what i am able to produce on XL 1.0:
The warning you're getting is likely caused by NANs showing up (e.g. divisions by zeros or something else) showing up somewhere during the process.
I get patchy results with it (sometimes it works sometimes it doesn't) It doesn't seem to work at all with SDXL models and only with certain sampler using 1.5. When I run it I get this output.
RuntimeWarning: invalid value encountered in cast img = Image.fromarray(np.clip(i, 0, 255).astype(np.uint8))
Any ideas what this might be?