meareabc

> 老哥解决了么 我也遇到了这种问题 折腾两天了 不知道你是否解决了问题，你可以参考这个方法试试看https://github.com/allenai/ai2thor/issues/993#issuecomment-1154744878

1.请更新最新版本的代码 2.下载预训练的embedding等模型 3.将 with torch.no_grad(): output= network(content,style) output = output.cpu() 更改为 with torch.no_grad(): output= network(content,style) output = output[0].cpu()

by using this config without deepseed ,will generate only one *.ckpt in output folder: python examples/wanvideo/train_wan_t2v.py \ --task train \ --train_architecture lora \ --dataset_path data/test \ --output_path ./output \ --dit_path...

maybe you should check your lora model "*.ckpt" should be a file but not a folder. if you use deepspeed, it may save as a folder

some time errors like this: [rank3]: Traceback (most recent call last): [rank3]: File "/home/lkh/sd/DiffSynth-Studio/examples/wanvideo/wan_14b_text_to_video_tensor_parallel.py", line 166, in [rank3]: trainer.test(model, dataloader) [rank3]: File "/home/lkh/anaconda3/envs/DiffSynth/lib/python3.10/site-packages/lightning/pytorch/trainer/trainer.py", line 775, in test [rank3]: return call._call_and_handle_interrupt(...