What's up with the 0.5 versions?
Do they add other types of cute bunnies?
just some snapshots we are testing with further training, so yes better bunnies but not worthy of 0.6 status
What’s the right way to prompt? I don’t see any difference from the official model.
What’s the right way to prompt? I don’t see any difference from the official model.
I have the same problem. Text2Video mode doesn't produce any noticeable results. I'm waiting for the beta version. It's probably a similar issue with the earlier version of the NSFW checkpoint for the LTX model, which also showed no differences from the original model.
What’s the right way to prompt? I don’t see any difference from the official model.
same. quite similar results, no additional abilities.
yep, youre right. thanks. theres an issue with the int8 models I posted, the conversion from bf16 to int8 that I did remove a large amount of the training data. working on it now, im hoping to post 0.6-beta which should fix it
yep, youre right. thanks. theres an issue with the int8 models I posted, the conversion from bf16 to int8 that I did remove a large amount of the training data. working on it now, im hoping to post 0.6-beta which should fix it
What made you decide to delete some of the data?
I didnt decide, I did it by accident. 0.6-beta which ill drop shortly will resolve it
beta-0.6 is uploaded, it should solve many of the issues you mention above. still training though, so there wil be a beta-0.7 too :)
beta-0.6 is uploaded, it should solve many of the issues you mention above. still training though, so there wil be a beta-0.7 too :)
I'm downloading this right now.