下载Youku-AliceMind的文件名与caption文件里的名字不同要怎么匹配？

[Science2AI-TaoXu]

[dai-yutong]

I'm facing the same issue. How are the filenames mapped?

[dai-yutong]

I'm facing the same issue. How are the filenames mapped?

I have solved my problem.

from datasets.utils.file_utils import hash_url_to_filename
url = f"public-unzip-dataset/modelscope/Youku-AliceMind/master/{filename_in_csv}"
new_filename = hash_url_to_filename(url)

[Sun-light-W]

I'm facing the same issue. How are the filenames mapped?

I have solved my problem.

from datasets.utils.file_utils import hash_url_to_filename url = f"public-unzip-dataset/modelscope/Youku-AliceMind/master/{filename_in_csv}" new_filename = hash_url_to_filename(url)

I used this method but the resulting new_filename was not found in my data_files directory

[dai-yutong]

I'm facing the same issue. How are the filenames mapped?

I have solved my problem.
from datasets.utils.file_utils import hash_url_to_filename url = f"public-unzip-dataset/modelscope/Youku-AliceMind/master/{filename_in_csv}" new_filename = hash_url_to_filename(url)

I used this method but the resulting new_filename was not found in my data_files directory

Maybe you can enter the directory of the modelscope package, find the file "msdatasets/utils/oss_utils.py", find the code filename = hash_url_to_filename(file_oss_key, etag=None), print file_oss_key and filename when you run MsDataset.load.