huggingface / huggingface/dataset-viewer
Dataset viewer gives error "Bad split" if the name of the split is not "train"
- Dominant language
- Python
- Stars
- 899
- Forks
- 130
- Avg merge
- 9h 26m
- Merged PRs (30d)
- 14
Description
Below is a screenshot of the Dataset viewer error I see after uploading this dataset: [https://huggingface.co/datasets/roemmele/SceneIllustrations](https://huggingface.co/datasets/roemmele/SceneIllustrations)
I've only been able to resolve this error by renaming the **test** split to **train** (which is an inaccurate name for my dataset since it consists of a test set only).
Also, here's the [discussion](https://discuss.huggingface.co/t/dataset-viewer-doesnt-work-if-split-train/168712) I previously opened regarding this, which includes some troubleshooting that ultimately didn't resolve the issue.
Contributor guide
Research direction
Start by reproducing the error with the linked SceneIllustrations dataset and its test split, then review the linked discussion for the existing troubleshooting context. Done means the Dataset viewer loads a dataset whose only split is not named train without reporting “Bad split”.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- api, backend, data
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 38/100