The data is not downloaded properly

#1
by SSEgpt - opened

Hi,

I tried to fine tune the Lerobot pi0 model with this databse, but the execution finished at 14% chunk-012 video file was not found, it seems like the data was not correctly donwloaded. In fact, I tried to download it manually by cloning it as a github repository, and it also crashed.

Here you have the error:

git clone https://huggingface.co/datasets/IPEC-COMMUNITY/fractal20220817_data_lerobot
Cloning into 'fractal20220817_data_lerobot'...
remote: Enumerating objects: 191126, done.
remote: Total 191126 (delta 0), reused 0 (delta 0), pack-reused 191126 (from 1)
Receiving objects: 100% (191126/191126), 129.76 MiB | 39.29 MiB/s, done.
Resolving deltas: 100% (1482/1482), done.
Updating files: 100% (174430/174430), done.
panic: runtime error: index out of range [0] with length 0KiB/s

goroutine 191 [running]:
github.com/git-lfs/git-lfs/tq.(*basicDownloadAdapter).download(0xc000011948, 0xc0079f0600, 0xc000742a40, 0x0, 0xc00bdf40b8, 0x0, 0xae4b40, 0xc00be06300, 0x0, 0x0)
github.com/git-lfs/git-lfs/tq/basic_download.go:161 +0x15d9
github.com/git-lfs/git-lfs/tq.(*basicDownloadAdapter).DoTransfer(0xc000011948, 0x0, 0x0, 0xc0079f0600, 0xc000742a40, 0x0, 0x0, 0x0)
github.com/git-lfs/git-lfs/tq/basic_download.go:101 +0x403
github.com/git-lfs/git-lfs/tq.(*adapterBase).worker(0xc0002a42d0, 0x6, 0x0, 0x0)
github.com/git-lfs/git-lfs/tq/adapterbase.go:182 +0x504
created by github.com/git-lfs/git-lfs/tq.(*adapterBase).Begin
github.com/git-lfs/git-lfs/tq/adapterbase.go:95 +0x255
fatal: the remote end hung up unexpectedly
warning: Clone succeeded, but checkout failed.
You can inspect what was checked out with 'git status'
and retry with 'git restore --source=HEAD :/'

IPEC at Shanghai AI Laboratory org

Try to remove the .cache dir and re-download again.

use the latest huggingface-clli

huggingface-cli download IPEC-COMMUNITY/fractal20220817_data_lerobot --repo-type dataset --local-dir fractal20220817_data_lerobot

I followed your steps and didn't work, it only donwloaded until video 12515, it looks like there is something wrong on the dataset, where it should be more than 87k videos. I tried downloading only the videos from a python script and the same occurred.

image.png

IPEC at Shanghai AI Laboratory org

could you please raise an issue in any4lerobot?

provide me with the script and the exception stack trace.

IPEC at Shanghai AI Laboratory org

please try pip install "huggingface_hub==0.32.2"

Tavish9 changed discussion status to closed

Sign up or log in to comment