tensorflow / tensorflow/datasets
Better Drive files download failure
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 4.6k
- Forks
- 1.6k
- Avg merge
- 3h 54m
- Merged PRs (30d)
- 1
Description
Download of drive urls sometimes fails with NonMatchingChecksumError: Artifact https://drive.google.com/... has wrong checksum.
Explanation: Drive sometimes reject the download attempt, and the rejection page is downloaded instead of the data:
- If the user is based in china (should use VPN)
- If there is too many downloads of the same file.
The best solution currently is to manually download the data (https://www.tensorflow.org/datasets/overview#manual_download_if_download_fails), rather than using the automated download which got rejected by drive.
Otherwise:
- Try the download latter on.
- Try on a different computer
- Rather than downloading the file in each colab connection, load the dataset from a GCS bucket. See instructions.
Not sure there can be a solution on Google Drive side, while preventing abuse.
On TFDS side, we could make the error message more explicit when we detect a drive URL.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by locating the Drive URL download and checksum-error handling in the TensorFlow Datasets codebase. Reproduce or inspect the reported NonMatchingChecksumError path, then determine where a Drive-specific explanation can be added. Done means rejected Drive downloads provide clearer guidance about manual downloads and retry options.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, tensorflow
- Domain
- data
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100