Rishabh Jain

Agree that from_single_file method will be helpful. Since animatediff was added to diffusers, new versions for SD as well as SDXL has been released. But unable to use them, as...

It's looking for only images inside `input folder` but jupyter has made an `.ipynb_checkpoints` directory there. Just remove this directory and it should work. `!rm -rf input/.ipythnb_checkpoints`

That's because your lora_rank is 1 here. Increasing it will increase the size. A rank of 4 will roughly give ~3MB.

You can take a look here: https://github.com/cloneofsimo/lora/blob/master/scripts/run_inference.ipynb on how to run inference with it.

You can have a look at some examples of good set of training params by looking at training scripts here. https://github.com/cloneofsimo/lora/tree/master/training_scripts

You can use the `--duration` flag, i.e. --duration 180

I have been running into it too. As an alternative, you can try using generated images for regularization instead of real ones.