How to load the dataset after git clone the repo manually?

#4
by cq - opened

Hi, I have met a problem when i try to load this dataset. As the download server of huggingface is not stable, I failed to load the dataset by directly using load_dataset. Therefore, I git clone this repo manually first. However, I still failed to load the dataset. Here are some snippets I tried.

from datasets import load_dataset
load_dataset("./diffusiondb", "2m_all") # download the dataset and ignore local files
load_dataset("./diffusiondb/images") # extract and load the dataset successfully, but the dataset lose the metadata, including prompts

from datasets import load_dataset_builder
db = load _dataset_builder("./diffusiondb/diffusiondb.py", "2m_all")
db.download_and_prepare() # still download the dataset and ignore local files

Could anyone help me to load the dataset using the pre-git-cloned repo files please?

Polo Club of Data Science org

Hi @cq , if you want to manually download the dataset, you can check out this downloader script:

xiaohk changed discussion status to closed

Sign up or log in to comment