Sam Sharpe

Results 2 comments of Sam Sharpe

Yes, we are currently using HF dataset in streaming mode since we have datasets that are 200GB on disk in parquet files. Trying to find a way to more efficiently...