There seem to be some problems with the code. For example, some.pth files cannot be found, and there are also strange path formats and so on.

[percy4]

else:
    print("\n---resuming from last.pth ckpt---\n")
    checkpoint = torch.load(
        f'/fs/scratch/PAS2490/mindeye/weights/datasets--pscotti--mindeyev2/snapshots/26421f100e4c6012a35ecadb272a0ec1d999202d/train_logs/final_subj01_pretrained_40sess_24bs/last.pth',
        map_location='cpu')
    model.load_state_dict(checkpoint['model_state_dict'], strict=False)
    del checkpoint

    ckpt = torch.load(f'/neuroclips/snapshots/183269ab73b49d2fa10b5bfe077194992934e4e6/sd_image_var_autoenc.pth')
    autoenc.load_state_dict(ckpt)

[gongzix]

Some pre-trained weight files come from the VAE of Stable Diffusion or other sources. Thank you for the reminder—this link (https://huggingface.co/datasets/pscotti/mindeyev2/tree/main) contains these files, and you can find the corresponding .pth files there.

Another reminder is that you might need to modify the paths of many files, as these files depend on what you download from Hugging Face or train yourself. The paths in the source code are for reference only.

[percy4]

Some pre-trained weight files come from the VAE of Stable Diffusion or other sources. Thank you for the reminder—this link (https://huggingface.co/datasets/pscotti/mindeyev2/tree/main) contains these files, and you can find the corresponding .pth files there.一些预训练的权重文件来自稳定扩散的VAE或其他来源。感谢您的提醒 - 此链接（https://huggingface.co/datasets/pscotti/mindeyev2/tree/main）包含这些文件，您可以在那里找到相应的 .pth 文件。

Another reminder is that you might need to modify the paths of many files, as these files depend on what you download from Hugging Face or train yourself. The paths in the source code are for reference only.另一个提醒是，您可能需要修改许多文件的路径，因为这些文件取决于您从 Hugging Face 下载的内容或自行训练的内容。源码中的路径仅供参考。

thanks！

[NicoleRR2001]

May I ask how can I find this pth