Datasets:
Dataset Viewer issue: TransformRowsProcessingError
The dataset viewer is not working on page 22 (it seems all other pages are working).
Error details:
Error code: TransformRowsProcessingError
This was brought to our attention in this GitHub issue (copied context below):
I'm trying to load this dataset with PyTorch dataloader, and got the following error.
SyntaxError: not a TIFF file (header b'IIU\x00\x18\x00\x00\x00' not valid)I downloaded the dataset from huggingface with the following code
from datasets import load_dataset ds = load_dataset("imageomics/rare-species")and used PyTorch Dataset and Dataloader to load it:
class CustomDataset(torch.utils.data.Dataset): def __init__(self, dataset, transform=None): self.dataset = dataset self.transform = transform def __len__(self): return len(self.dataset) def __getitem__(self, idx): return self.dataset[idx]["rarespecies_id"] custom_dataset = CustomDataset(dataset["train"], transform=transform) dataloader = DataLoader(custom_dataset, batch_size=32, shuffle=False) for i, data in enumerate(dataloader): print(data)CustomDataset returns
rarespecies_idonly for simplicity for debugging. I was able to get the ID's printed for the first 66 batches and got the SyntaxErrorSyntaxError: not a TIFF file (header b'IIU\x00\x18\x00\x00\x00' not valid)when loading the 67th batch.
Hi, is there anything that can be done with this @albertvillanova @lhoestq @severo ?
I will add that we did re-download all the images and haven't found anything corrupted; also, they're all JPGs. Any suggestions for other avenues we could explore would be appreciated.
It's in row 2169: https://datasets-server.huggingface.co/rows?dataset=imageomics%2Frare-species&config=default&split=train&offset=2169&length=1 -> gives the error {"error":"Server error while post-processing the split rows. Please report the issue."}
I assume it's this row in metadata.csv:
f5332685-0953-4623-a2fd-1e0229e64eea,13989194,858738,Animalia,Arthropoda,Insecta,Orthoptera,Tettigoniidae,Saga,pedo,Saga pedo,Predatory Bush-cricket,dataset/Animalia-Arthropoda-Insecta-Orthoptera-Tettigoniidae-Saga-pedo/13989194_858738_eol-full-size-copy.jpg
yet the image does not seem wrong: https://huggingface.co/datasets/imageomics/rare-species/blob/main/dataset/Animalia-Arthropoda-Insecta-Orthoptera-Tettigoniidae-Saga-pedo/13989194_858738_eol-full-size-copy.jpg
I'm recomputing the parquet files, in case we had some issue during the conversion...
During the refresh of step config-parquet-and-info, I get this error: Permission denied: '/tmp/hf-datasets-cache/medium/datasets/62628918354390-config-parquet-and-info-imageomics-rare-species-dc40b4ef/downloads/db26be421fcb817885acfbe6fa6aefdf2670894421d0303e8c41377402328d2e.incomplete'. I'm not sure why we have this error. cc
@polinaeterna
or
@lhoestq
That is so odd. Thank you for trying the recompute! Please let us know if there's anything else we can do on our end.
We're working on it (https://github.com/huggingface/datasets/pull/7105). We will recompute the step automatically when the fix is deployed.
Hi ! Not yet ^^' we recently released datasets 3.0 and we will this is version for the viewer
Thanks for clarifying!
This problem still exists with datasets 3.0.1. Any update?
The bug has been fixed ! 🤗
