d2l-ai/d2l-zh · error · AssertionError

f"{name} 不存在于 {DATA_HUB}"

Error message

f"{name} 不存在于 {DATA_HUB}"

What it means

An AssertionError in d2l.tensorflow.download stating that the requested dataset name is not a key in DATA_HUB. DATA_HUB is a module-level dict mapping dataset names to (url, sha1) pairs; download() refuses to fetch anything not registered there because it needs the pinned URL and checksum. The Chinese message reads '{name} does not exist in {DATA_HUB}'.

Source

Thrown at d2l/tensorflow.py:363

def evaluate_loss(net, data_iter, loss):
    """评估给定数据集上模型的损失

    Defined in :numref:`sec_model_selection`"""
    metric = d2l.Accumulator(2)  # 损失的总和,样本数量
    for X, y in data_iter:
        l = loss(net(X), y)
        metric.add(d2l.reduce_sum(l), d2l.size(l))
    return metric[0] / metric[1]

DATA_HUB = dict()
DATA_URL = 'http://d2l-data.s3-accelerate.amazonaws.com/'

def download(name, cache_dir=os.path.join('..', 'data')):
    """下载一个DATA_HUB中的文件，返回本地文件名

    Defined in :numref:`sec_kaggle_house`"""
    assert name in DATA_HUB, f"{name} 不存在于 {DATA_HUB}"
    url, sha1_hash = DATA_HUB[name]
    os.makedirs(cache_dir, exist_ok=True)
    fname = os.path.join(cache_dir, url.split('/')[-1])
    if os.path.exists(fname):
        sha1 = hashlib.sha1()
        with open(fname, 'rb') as f:
            while True:
                data = f.read(1048576)
                if not data:
                    break
                sha1.update(data)
        if sha1.hexdigest() == sha1_hash:
            return fname  # 命中缓存
    print(f'正在从{url}下载{fname}...')
    r = requests.get(url, stream=True, verify=True)
    with open(fname, 'wb') as f:
        f.write(r.content)
    return fname

View on GitHub (pinned to e6b18ccea7)

Solutions

1. Register the dataset before downloading: DATA_HUB['my_dataset'] = (DATA_URL + 'my_file.csv', '<sha1>'); download('my_dataset').
2. Use one of the pre-registered names at the bottom of the module, e.g. d2l.download('kaggle_house_train') or 'kaggle_house_test'.
3. For real Kaggle competitions, download manually from kaggle.com and place files in the ../data cache_dir; download() only serves the D2L mirror.
4. Compute the SHA1 with: python -c "import hashlib;print(hashlib.sha1(open('file','rb').read()).hexdigest())" when registering a custom entry.

Example fix

# before
d2l.download('kaggle_house_prediction')  # AssertionError: not in DATA_HUB
# after
d2l.DATA_HUB['kaggle_house_train'] = (
    d2l.DATA_URL + 'kaggle_house_pred_train.csv',
    '020e2b8f8f8c6f6f6f6f6f6f6f6f6f6f6f6f6f6f')
d2l.download('kaggle_house_train')

def safe_download(name):
    if name not in d2l.DATA_HUB:
        available = ', '.join(sorted(d2l.DATA_HUB))
        raise KeyError(f'{name!r} not registered. Available: {available}')
    return d2l.download(name)

def is_registered(name: str) -> bool:
    return isinstance(name, str) and name in d2l.DATA_HUB

try:
    d2l.download(name)
except AssertionError:
    raise KeyError(f'{name!r} not in DATA_HUB; register it or pick from {sorted(d2l.DATA_HUB)}')

When it happens

Trigger: Calling d2l.download('my_dataset') or download_extract/download_all with a name never registered; requesting 'kaggle_house_train' from a fresh interpreter where the DATA_HUB[...] assignment lines at module bottom were not executed (e.g. partially imported module); typos in the name string.

Common situations: Users assuming d2l can download arbitrary Kaggle files (the book's sec_kaggle_house registers only the mirrored kaggle_house_train/kaggle_house_test CSVs on d2l-data.s3); copying download() into a notebook but not the DATA_HUB registration lines; name case or underscore mismatches like 'kaggle-house-train'.

Related errors

• f"{name} 不存在于 {DATA_HUB}"
• f"{name} 不存在于 {DATA_HUB}"
• f"{name} 不存在于 {DATA_HUB}"
• 只有zip/tar文件可以被解压缩
• 只有zip/tar文件可以被解压缩

AI-assisted analysis of d2l-ai/d2l-zh@e6b18ccea7 (2026-08-14). Data as JSON: /api/errors/b3c9bda87e44390c. Report an issue: GitHub.