Progression feedback on Beam related processing?
Hi,
I'm wondering if there could be a way to add some progression indication about the Beam processing step, since it looks like it can take a lot of time.
Thanks!
Hi ! It should be possible by checking the examples Counter metrics of the Apache Beam pipeline, instead of calling pipeline_results.wait_until_finish() directly in builder.py in the datasets code. Is this something you'd be interested in contributing ? I'd be happy to give some pointers. Otherwise, feel free to to open an issue at https://github.com/huggingface/datasets
I'd be interested in trying to tackle this! Would you have a preference in terms of display (printing status from Beam vs. integrating in a fancier tqdm-like progress bar)?
Cool ! I think a tqdm bar would be ideal, similarly to what we use in GeneratorBasedBuilder._prepare_split. Though in general we don't know how many examples there are in advance, so it will just show the number of examples growing ^^'
Hi @uj .
Note that we are going to deprecate this dataset: it only contains the pre-processed data for 6 of the languages.
I would recommend that you use the current official "wikimedia/wikipedia" dataset, with the pre-processed data for all the languages and for the latest dump 2023-11-01.