MCPcopy Create free account
hub / github.com/FunAudioLLM/ThinkSound / get_all_s3_urls

Function get_all_s3_urls

ThinkSound/data/dataset.py:850–897  ·  view source on GitHub ↗

get urls of shards (tar files) for multiple datasets in one s3 bucket

(
    names=[],           # list of all valid [LAION AudioDataset] dataset names
    # list of subsets you want from those datasets, e.g. ['train','valid']
    subsets=[''],
    s3_url_prefix=None,  # prefix for those dataset names
    recursive=True,     # recursively list all tar files in all subdirs
    filter_str='tar',   # only grab files with this substring
    # print debugging info -- note: info displayed likely to change at dev's whims
    debug=False,
    profiles={},        # dictionary of profiles for each item in names, e.g. {'dataset1': 'profile1', 'dataset2': 'profile2'}
)

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

Callers 1

load_data_urlsMethod · 0.85

Calls 1

get_s3_contentsFunction · 0.85

Tested by

no test coverage detected