Datasets:
Tasks:
Text Generation
Modalities:
Text
Sub-tasks:
language-modeling
Languages:
English
Size:
100K - 1M
License:
File size: 654 Bytes
231bd79 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 |
from datasets import load_dataset
from itertools import islice
import sys
import time
from tqdm import tqdm
dataset = load_dataset("./proof-pile.py", "default")
size = dataset["train"].dataset_size / 2**30
print(f"{size} GB TRAIN TOTAL")
print(dataset)
for x in tqdm(dataset["train"]):
print("EXAMPLE INSTANCE (trimmed):")
print(x["text"][:100])
break
then = time.time()
for x in tqdm(dataset["train"]):
pass
now = time.time()
print(f"{size} GB TRAIN TOTAL")
print(f"TRAVERSED IN {now-then} SECONDS")
size += dataset["validation"].dataset_size/2**30 + dataset["test"].dataset_size/2**30
print(f"{size} GB TOTAL (TRAIN, VAL, TEST)")
|