Datasets:
Tasks:
Audio Classification
Modalities:
Audio
Languages:
English
Size:
10K<n<100K
Tags:
audio
License:
Parsing error
#1
by
RaphaelSchwinger
- opened
Hey, thanks for providing this dataset.
When I try to load it with:
import datasets
audioset = datasets.load_dataset('agkphysics/AudioSet')
I get the following parsing error:
Failed to read file '/workspaces/COLA-light/data/downloads/7b8b212193e37ba5a7e3b20778d6540d1fa914ba74d976b87c50887f3c7f7946' with error <class 'pandas.errors.ParserError'>: Error tokenizing data. C error: Expected 4 fields in line 4, saw 5
ERROR:datasets.packaged_modules.csv.csv:Failed to read file '/workspaces/COLA-light/data/downloads/7b8b212193e37ba5a7e3b20778d6540d1fa914ba74d976b87c50887f3c7f7946' with error <class 'pandas.errors.ParserError'>: Error tokenizing data. C error: Expected 4 fields in line 4, saw 5
Any ideas, what's wrong?
Greetings
It's trying to load from the CSV files but they're not in the correct format.
I intend to create a custom loading script soon so that it can be used with the datasets
library.
@RaphaelSchwinger
I have uploaded a dataset loading script, so you should now be able to download/stream the data using load_dataset()
.
agkphysics
changed discussion status to
closed