vadim epstein

training failure - how to fix it?

5

we have attempted to reproduce your results by training the models from scratch with provided data and settings (`pirounet_dance`, `pirounet_watch`, and the commented hyperparams in the `default_config.py`). unfortunately, we didn't...

nothing is plotted on evaluation

3

hello, thanks for opensourcing this intriguing work! it seems some breaking changes happened in the dependencies though, as nothing is plotted when running `main_eval.py`. i see some data generated in...

MotionAdapter.from_single_file

16

### Model/Pipeline/Scheduler description `Animatediff` is a great technique with exploding interest and advancements; additional custom models start popping around in `ckpt` format, which cannot be loaded to `diffusers` in the...

not compatible with diffusers 0.21+ [with workaround]

2

everything runs ok on diffusers version 0.20 or below, while getting this error on diffusers 0.21: ``` File "F:\_neuro\SDfu\lib\tokenflow.py", line 185, in denoise_step noise_pred = self.unet(lat_in, t, conds).sample File "C:\Users\eps\AppData\Local\Programs\Python\Python39\lib\site-packages\torch\nn\modules\module.py",...

for those who (like me) wanted to apply this exciting technique for longer videos: i've integrated this method into my SD repo https://github.com/eps696/SDfu and added there batches for pivots with...