>>> from datasets import load_dataset
>>> dataset = load_dataset('HuggingFaceFW/fineweb', split='train', streaming=True)
>>> print(next(iter(dataset)))
{'text': 'How AP reported in all formats from tornado-stricken regionsMarch 8, 2012\nWhen the first serious bout of tornadoes of 2012 blew through middle America in the middle of the night, they touched down in places hours from any AP bureau...', ...,
 'language_score': 0.9721424579620361, 'token_count': 717}

>>> from datasets import load_dataset
>>> data_files = {'train': 'path/to/OSCAR-2201/compressed/en_meta/*.jsonl.gz'}
>>> dataset = load_dataset('json', data_files=data_files, split='train', streaming=True)
>>> print(next(iter(dataset)))
{'id': 0, 'text': 'Founded in 2015, Golden Bees is a leading programmatic recruitment platform dedicated to employers, HR agencies and job boards. The company has developed unique HR-custom technologies and predictive algorithms to identify and attract the best candidates for a job opportunity.', ...

>>> from datasets import load_dataset
>>> dataset = load_dataset('HuggingFaceFW/fineweb', split='train', streaming=True, columns=["url", "date"])
>>> print(next(iter(dataset)))
{'url': 'http://%20jwashington@ap.org/Content/Press-Release/2012/How-AP-reported-in-all-formats-from-tornado-stricken-regions', 'date': '2013-05-18T05:48:54Z'}
>>> dataset = load_dataset('HuggingFaceFW/fineweb', split='train', streaming=True, filters=[("language_score", ">=", 0.99)])
>>> print(next(iter(dataset)))
{'text': 'Everyone wishes for something. And lots of people believe they know how to make their wishes come true with magical thinking.\nWhat is it? "Magical thinking is a belief in forms of causation, with no known physical basis," said Professor Emily Pronin of Princeton...', ...,
 'language_score': 0.9900368452072144, 'token_count': 716}

>>> from datasets import load_dataset
>>> dataset = load_dataset("allenai/c4", "en", streaming=True, split="train")
>>> print(next(iter(dataset["text"])))
Beginners BBQ Class Taking Place in Missoula!...

>>> from datasets import load_dataset

# faster 🐇
>>> dataset = load_dataset("ethz/food101")
>>> iterable_dataset = dataset.to_iterable_dataset()

# slower 🐢
>>> iterable_dataset = load_dataset("ethz/food101", streaming=True)

>>> import torch
>>> from datasets import load_dataset

>>> dataset = load_dataset("ethz/food101")
>>> iterable_dataset = dataset.to_iterable_dataset(num_shards=64) # shard the dataset
>>> iterable_dataset = iterable_dataset.shuffle(buffer_size=10_000)  # shuffles the shards order and use a shuffle buffer when you start iterating
dataloader = torch.utils.data.DataLoader(iterable_dataset, num_workers=4)  # assigns 64 / 4 = 16 shards from the shuffled list of shards to each worker when you start iterating

>>> from datasets import load_dataset
>>> dataset = load_dataset('HuggingFaceFW/fineweb', split='train', streaming=True)
>>> shuffled_dataset = dataset.shuffle(seed=42, buffer_size=10_000)

>>> for epoch in range(epochs):
...     shuffled_dataset.set_epoch(epoch)
...     for example in shuffled_dataset:
...         ...

>>> dataset = load_dataset('HuggingFaceFW/fineweb', split='train', streaming=True)
>>> dataset_head = dataset.take(2)
>>> list(dataset_head)
[{'text': "How AP reported in all formats from tor...},
 {'text': 'Did you know you have two little yellow...}]

>>> train_dataset = shuffled_dataset.skip(1000)

>>> from datasets import load_dataset
>>> dataset = load_dataset("fancyzhx/amazon_polarity", split="train", streaming=True)
>>> print(dataset)
IterableDataset({
    features: ['label', 'title', 'content'],
    num_shards: 4
})

>>> dataset.shard(num_shards=2, index=0)
IterableDataset({
    features: ['label', 'title', 'content'],
    num_shards: 2
})

>>> dataset.reshard()
IterableDataset({
    features: ['label', 'title', 'content'],
    num_shards: 3600
})

>>> from datasets import concatenate_datasets, load_dataset

>>> stories = load_dataset("ajibawa-2023/General-Stories-Collection", split="train", streaming=True)
>>> stories = stories.select_columns(["text"])  # only keep the 'text' column
IterableDataset({
    features: Unknown,
    num_shards: 10
})

>>> wiki = load_dataset("wikimedia/wikipedia", "20231101.en", split="train", streaming=True)
>>> wiki = wiki.select_columns(["text"])  # only keep the 'text' column
IterableDataset({
    features: ['text'],
    num_shards: 41
})

>>> bert_dataset = concatenate_datasets([stories, wiki])
IterableDataset({
    features: ['text'],
    num_shards: 51
})

>>> from datasets import IterableDataset
>>> stories_ids = IterableDataset.from_dict({"ids": list(range(num_stories))})
>>> stories_with_ids = concatenate_datasets([stories, stories_ids], axis=1)

>>> from datasets import interleave_datasets
>>> es_dataset = load_dataset('allenai/c4', 'es', split='train', streaming=True)
IterableDataset({
    features: Unknown,
    num_shards: 2048
})
>>> fr_dataset = load_dataset('allenai/c4', 'fr', split='train', streaming=True)
IterableDataset({
    features: Unknown,
    num_shards: 2048
})

>>> multilingual_dataset = interleave_datasets([es_dataset, fr_dataset])
IterableDataset({
    features: ['text', 'timestamp', 'url'],
    num_shards: 2048
})
>>> list(multilingual_dataset.take(2))
[{'text': 'Comprar Zapatillas para niña en chancla con goma por...'},
 {'text': 'Le sacre de philippe ier, 23 mai 1059 - Compte Rendu...'}]

>>> multilingual_dataset_with_oversampling = interleave_datasets([es_dataset, fr_dataset], probabilities=[0.8, 0.2], seed=42)
>>> list(multilingual_dataset_with_oversampling.take(2))
[{'text': 'Comprar Zapatillas para niña en chancla con goma por...'},
 {'text': 'Chevrolet Cavalier Usados en Bogota - Carros en Vent...'}]

>>> from datasets import load_dataset
>>> dataset = load_dataset('allenai/c4', 'en', streaming=True, split='train')
>>> dataset = dataset.rename_column("text", "content")

>>> from datasets import load_dataset
>>> dataset = load_dataset('allenai/c4', 'en', streaming=True, split='train')
>>> dataset = dataset.remove_columns('timestamp')

>>> from datasets import load_dataset
>>> dataset = load_dataset('nyu-mll/glue', 'mrpc', split='train', streaming=True)
>>> dataset.features
{'sentence1': Value('string'),
'sentence2': Value('string'),
'label': ClassLabel(names=['not_equivalent', 'equivalent']),
'idx': Value('int32')}

>>> from datasets import ClassLabel, Value
>>> new_features = dataset.features.copy()
>>> new_features["label"] = ClassLabel(names=['negative', 'positive'])
>>> new_features["idx"] = Value('int64')
>>> dataset = dataset.cast(new_features)
>>> dataset.features
{'sentence1': Value('string'),
'sentence2': Value('string'),
'label': ClassLabel(names=['negative', 'positive']),
'idx': Value('int64')}

>>> dataset.features
{'audio': Audio(sampling_rate=44100, mono=True)}

>>> dataset = dataset.cast_column("audio", Audio(sampling_rate=16000))
>>> dataset.features
{'audio': Audio(sampling_rate=16000, mono=True)}

>>> def add_prefix(example):
...     example['text'] = 'My text: ' + example['text']
...     return example

>>> from datasets import load_dataset
>>> dataset = load_dataset('allenai/c4', 'en', streaming=True, split='train')
>>> updated_dataset = dataset.map(add_prefix)
>>> list(updated_dataset.take(3))
[{'text': 'My text: Beginners BBQ Class Taking Place in Missoula!\nDo you want to get better at making...',
  'timestamp': '2019-04-25 12:57:54',
  'url': 'https://klyq.com/beginners-bbq-class-taking-place-in-missoula/'},
 {'text': 'My text: Discussion in \'Mac OS X Lion (10.7)\' started by axboi87, Jan 20, 2012.\nI\'ve go...',
  'timestamp': '2019-04-21 10:07:13',
  'url': 'https://forums.macrumors.com/threads/restore-from-larger-disk-to-smaller-disk.1311329/'},
 {'text': 'My text: Foil plaid lycra and spandex shortall with metallic slinky insets. Attached metall...',
  'timestamp': '2019-04-25 10:40:23',
  'url': 'https://awishcometrue.com/Catalogs/Clearance/Tweens/V1960-Find-A-Way'}]

>>> updated_dataset = dataset.map(add_prefix, remove_columns=["timestamp", "url"])
>>> list(updated_dataset.take(3))
[{'text': 'My text: Beginners BBQ Class Taking Place in Missoula!\nDo you want to get better at making...'},
 {'text': 'My text: Discussion in \'Mac OS X Lion (10.7)\' started by axboi87, Jan 20, 2012.\nI\'ve go...'},
 {'text': 'My text: Foil plaid lycra and spandex shortall with metallic slinky insets. Attached metall...'}]

>>> from datasets import load_dataset
>>> from transformers import AutoTokenizer
>>> dataset = load_dataset("allenai/c4", "en", streaming=True, split="train")
>>> tokenizer = AutoTokenizer.from_pretrained('distilbert-base-uncased')
>>> def encode(examples):
...     return tokenizer(examples['text'], truncation=True, padding='max_length')
>>> dataset = dataset.map(encode, batched=True, remove_columns=["text", "timestamp", "url"])
>>> next(iter(dataset))
{'input_ids': [101, 4088, 16912, 22861, 4160, 2465, 2635, 2173, 1999, 3335, ..., 0, 0, 0],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ..., 0, 0]}

>>> from datasets import load_dataset
>>> dataset = load_dataset('HuggingFaceFW/fineweb', streaming=True, split='train')
>>> start_with_ar = dataset.filter(lambda example: example['text'].startswith('San Francisco'))
>>> next(iter(start_with_ar))
{'text': 'San Francisco 49ers cornerback Shawntae Spencer will miss the rest of the sea...}

>>> even_dataset = dataset.filter(lambda example, idx: idx % 2 == 0, with_indices=True)
>>> list(even_dataset.take(3))
[{'text': 'How AP reported in all formats from tornado-stricken regionsMarch 8, 2012 Whe...},
 {'text': 'Car Wash For Clara! Now is your chance to help! 2 year old Clara Woodward has...},
 {'text': 'Log In Please enter your ECode to log in. Forgotten your eCode? If you create...}]

from datasets import load_dataset

# Load a dataset in streaming mode
dataset = load_dataset("some_dataset", split="train", streaming=True)

# Create batches of 32 samples
batched_dataset = dataset.batch(batch_size=32)

# Iterate over the batched dataset
for batch in batched_dataset:
    print(batch)
    break

batched_dataset = dataset.batch(batch_size=32, drop_last_batch=True)

>>> iterable_dataset = Dataset.from_dict({"a": range(6)}).to_iterable_dataset(num_shards=3)
>>> for idx, example in enumerate(iterable_dataset):
...     print(example)
...     if idx == 2:
...         state_dict = iterable_dataset.state_dict()
...         print("checkpoint")
...         break
>>> iterable_dataset.load_state_dict(state_dict)
>>> print(f"restart from checkpoint")
>>> for example in iterable_dataset:
...     print(example)

{'a': 0}
{'a': 1}
{'a': 2}
checkpoint
restart from checkpoint
{'a': 3}
{'a': 4}
{'a': 5}

>>> from torchdata.stateful_dataloader import StatefulDataLoader
>>> iterable_dataset = load_dataset("deepmind/code_contests", streaming=True, split="train")
>>> dataloader = StatefulDataLoader(iterable_dataset, batch_size=32, num_workers=4)
>>> # checkpoint
>>> state_dict = dataloader.state_dict()  # uses iterable_dataset.state_dict() under the hood
>>> # resume from checkpoint
>>> dataloader.load_state_dict(state_dict)  # uses iterable_dataset.load_state_dict() under the hood

dataset.push_to_hub("username/my_dataset")

dataset.push_to_hub("username/my_dataset", num_proc=8)

from datasets import load_dataset
reloaded_dataset = load_dataset("username/my_dataset")

>>> dataset.to_csv("path/of/my/dataset.csv")

>>> num_shards = dataset.num_shards
>>> for index in range(num_shards):
...     shard = dataset.shard(index, num_shards)
...     shard.to_parquet(f"path/of/my/dataset/data-{index:05d}.parquet")