Lightricks
Generated video are noisy and pixelated
I've an issue with the model on my Apple Silicon M1. I run it using the ComfyUI Integration.
For all the tests I've used this example prompt:
A woman with long brown hair and light skin smiles at another woman with long blonde hair. The woman with brown hair wears a black jacket and has a small, barely noticeable mole on her right cheek. The camera angle is a close-up, focused on the woman with brown hair's face. The lighting is warm and natural, likely from the setting sun, casting a soft glow on the scene. The scene appears to be real-life footage.
All the generated videos appear as random moving pixels. I used 30 steps generation, 768x512 size, batch size=1, cfg=3, max shift=2.15, base shift=0.95
Here's an example of the output. All videos that I've generated are pretty similar
https://github.com/user-attachments/assets/3ee32ae0-38d3-4320-80b0-660a8324f929
Generation is slow but it run without any error
Requested to load VideoVAE loaded completely 9.5367431640625e+25 799.548433303833 True Prompt executed in 102179.12 seconds Requested to load LTXV loaded completely 9.5367431640625e+25 3667.902587890625 True 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████| 30/30 [26:15<00:00, 52.51s/it] Prompt executed in 1686.05 seconds
Thanks in advance for the help
Any resolution on this? I'm seeing the same symptom with the same prompt (Apple M4 Max).
What version of PyTorch are you using? It works well on my Mac with PyTorch 2.3
