ruolinsss

why the output of waterz.agglomerate is a generator? BTW, is this package used in MALA project to obtain segmentations from pred_affinities?

2张H20，单卡可以跑推理，但用`accelerate launch --multi_gpu --num_processes 2 examples/wanvideo/model_inference/Wan2.2-I2V-A14B.py` 跑推理会oom，且`nvidia-smi`第二张卡一直没看到占用，这个是正确的启动单卡推理的方式吗？

运行其他wan2.2相关模型都可以在H20上顺利推理，但vace模型会在运行完50步diffusion后oom，调小resolution和frame number并没有帮助

Hi all, I’m encountering OOM issues when using LoRA to finetune Wan2.2-I2V-A14B with the script `lora/Wan2.2-I2V-A14B.sh` on 2×80 GB H20 GPUs. Is this expected? I noticed in the documentation that...

How to generate a longer video?

1

Hi, I encounter the problem with video duration, when i use follow instruction, the generated video is always 5s, how could I generate a longer video? `python vace/vace_pipeline.py --base wan...