Datasets:

Languages:
English
ArXiv:

how do I download only one source?

#9
by ishapuri - opened

Say I only want to download the data from Books or Wikipedia. How do I do that?

train_10_20_ds = datasets.load_dataset('bookcorpus', split='train[10:20]')

#!/bin/bash

Initialize the variable i

for ((i=1; i<=500; i++))
do
# Construct the URL with the current value of i
url="https://huggingface.co/datasets/cerebras/SlimPajama-627B/resolve/main/train/chunk1/example_train_${i}.jsonl.zst"

# Use wget to download the file
wget -p slimpajama_sample "$url"

done

#!/bin/bash

Initialize the variable i

for ((i=1; i<=500; i++))
do
url="https://huggingface.co/datasets/cerebras/SlimPajama-627B/resolve/main/train/chunk1/example_train_${i}.jsonl.zst"
wget -p slimpajama_sample "$url"
done

Entry not found

#!/bin/bash
cd /download_zstd
Initialize the variable i
for ((i=1; i<=500; i++))
do
url="https://huggingface.co/datasets/cerebras/SlimPajama-627B/resolve/main/train/chunk1/example_train_${i}.jsonl.zst"
wget -p slimpajama_sample "$url"
done

cd /download_zstd/huggingface.co/datasets/cerebras/SlimPajama-627B/resolve/main/train/chunk1
zstd -d *.zst
mkdir /mnt/hdd/download_zstd/jsonls/
mv *.jsonl /mnt/hdd/download_zstd/jsonls/

Sign up or log in to comment