Datasets:
BAAI
/

ArXiv:

No content in data

#3
by airlsyn - opened

I download some data from ModelScopes BAAI/CCI4.0-M2-Base-v1 (https://www.modelscope.cn/datasets/BAAI/CCI4.0-M2-Base-v1). But there is no content in the jsonl files, only meta info, like:

BAAI/CCI4.0-M2-Base-v1/Nemotron-CC-high-actual-actual-high/026_00005.jsonl

{"metadata":{"language":"eng","warc_record_id":"28fcb33c-8466-433c-9f33-f2e45bfbc90f","url":"https:\/\/destinypapers.net\/louis-pasteurs-experiment-that-refuted-spontaneous-generation-and-robert-kochs-development-of-kochs-postulates-were-both-critical-advances-in-the-field-of-microbiology-pick\/","quality":{"fine_web":2.5753555297851562,"fine_web_normal":2.5753555297851562},"domain":"Science","loss":1.6638370752334595,"toxicity":0.0005100775160826743}}

BAAI/CCI4.0-M2-Base-v1/Nemotron-CC-high-actual-actual-low/005_00005.jsonl

{"metadata":{"language":"eng","warc_record_id":"15763a88-8751-4950-8c8c-618b6fddc548","url":"http:\/\/www.seventeen.com\/entertainment\/justin-bieber-copy-cats-danny-avila","quality":{"fine_web":0.4193127751350403,"fine_web_normal":0.4193127751350403},"domain":"Arts_and_Entertainment","loss":2.9044601917266846,"toxicity":0.0005132229416631162}}

So how to get the real content of this data? Thanks a lot.

Beijing Academy of Artificial Intelligence org

The CCI4.0-M2-Base-v1 Chinese-English CC dataset is derived from the Nemotron-CC dataset, with some filtering applied.
For licence reasons, we have not included the original content of the samples. You can use the "warc_record_id" and "url" in the samples to match and retrieve the original content from the Nemotron-CC dataset.
For example, the subdirectory Nemotron-CC-high-actual-actual-* corresponds to the subdirectory Nemotron-CC-high-actual-actual in the Nemotron-CC dataset.

The CCI4.0-M2-Base-v1 Chinese-English CC dataset is derived from the Nemotron-CC dataset, with some filtering applied.
For licence reasons, we have not included the original content of the samples. You can use the "warc_record_id" and "url" in the samples to match and retrieve the original content from the Nemotron-CC dataset.
For example, the subdirectory Nemotron-CC-high-actual-actual-* corresponds to the subdirectory Nemotron-CC-high-actual-actual in the Nemotron-CC dataset.

Thanks a lot.

ldwang changed discussion status to closed

Sign up or log in to comment