在运行sh scripts/promptkd/base2new_train.sh imagenet 3时，显示报错如下，test数据加载不进来，train数据可以加载进来，恳求解决方案，谢谢

[XiaozhenQiao]

Traceback (most recent call last):
File "/nvfile-data/qxz/project/VQA/PromptKD/train.py", line 273, in
main(args)
File "/nvfile-data/qxz/project/VQA/PromptKD/train.py", line 204, in main
trainer = build_trainer(cfg)
File "/nvfile-data/qxz/project/VQA/PromptKD/Dassl/dassl/engine/build.py", line 11, in build_trainer
return TRAINER_REGISTRY.get(cfg.TRAINER.NAME)(cfg)
File "/nvfile-data/qxz/project/VQA/PromptKD/Dassl/dassl/engine/trainer.py", line 324, in init
self.build_data_loader()
File "/nvfile-data/qxz/project/VQA/PromptKD/Dassl/dassl/engine/trainer.py", line 347, in build_data_loader
dm = DataManager(self.cfg)
File "/nvfile-data/qxz/project/VQA/PromptKD/Dassl/dassl/data/data_manager.py", line 130, in init
test_loader = build_data_loader(
File "/nvfile-data/qxz/project/VQA/PromptKD/Dassl/dassl/data/data_manager.py", line 44, in build_data_loader
assert len(data_loader) > 0
AssertionError

[zhengli97]

你这个看起来应该是路径或者文件没放对，触发了assert len(data_loader)>0，就是你自己数据集存放的问题，没读起来

[XiaozhenQiao]

[XiaozhenQiao]

您好，我的这个文件放置感觉没问题呀，train和val在一级目录下，而且能够读取train数据集呀

[XiaozhenQiao]

val文件夹下面是50000张图片

[zhengli97]

val应该是这样的

[XiaozhenQiao]

好嘞，谢谢您的耐心回复

[zhengli97]

@XiaozhenQiao 最近我跑imagenet的实现，发现了一个对你现在情况的一个解决办法，你可以看一下这个，用valprep.sh脚本来处理文件，把val文件都放进文件夹里 https://github.com/soumith/imagenetloader.torch/blob/master/valprep.sh