google-deepmind / open_x_embodiment

Apache License 2.0
794 stars 54 forks source link

Not able to download or load certain data #5

Open youliangtan opened 11 months ago

youliangtan commented 11 months ago

Hi, I have some issue downloading the data.

Here, I followed the colab notebook to load the data, yet getting some errors in the notebook.

for dataset_name in tqdm.tqdm(DATASET_NAMES):
  _ = tfds.load(dataset_name, data_dir=DOWNLOAD_DIR)

Then, i just follow the readme to directly download the data using gsutil

gsutil -m cp -r gs://gdm-robotics-open-x-embodiment/{dataset_name} ~/tensorflow_datasets/

This mostly gives me empty data. Changing the bucket dir path to gs://gresearch/robotics/ works.

AasherH commented 7 months ago

Hi, I am running gsutil -m cp -r gs://gresearch/robotics/{dataset_name} ~/tensorflow_datasets/, where {dataset_name} = "fractal_20220817_data" for instance. However, I repeatedly get an error: 1 file/object could not be transferred. Have you found a workaround for this?

kpertsch commented 7 months ago

You have one underscore too many in your dataset name gs://gresearch/robotics/fractal20220817_data

Just tested that gsutil -m cp -r gs://gresearch/robotics/fractal20220817_data ~/tensorflow_datasets/ is working

AasherH commented 7 months ago

Thanks, that fixed it. However, I note that the colab tutorial includes the extra underscore, which could cause errors for other users.

kpertsch commented 7 months ago

The most up-to-date colab linked on the website seems to have the correct dataset name: https://colab.research.google.com/github/google-deepmind/open_x_embodiment/blob/main/colabs/Open_X_Embodiment_Datasets.ipynb

Maybe you're using an out-dated colab?

AasherH commented 7 months ago

Its later on in the colab document which has the extra underscore.

kpertsch commented 7 months ago

Ah my bad -- thanks for the catch! @quanvuong could you fix it in the Colab?