five

usvsnsp/generation-semantic-filters

收藏
Hugging Face2024-02-26 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/usvsnsp/generation-semantic-filters
下载链接
链接失效反馈
官方服务:
资源简介:
--- dataset_info: features: - name: sequence_id dtype: int64 - name: tokens sequence: int64 - name: text dtype: string - name: is_incrementing dtype: bool - name: is_repeating dtype: bool - name: sequence_duplicates dtype: int64 - name: max_frequency dtype: int64 - name: avg_frequency dtype: float64 - name: min_frequency dtype: int64 - name: median_frequency dtype: float64 - name: p25_frequency dtype: int64 - name: p75_frequency dtype: int64 - name: frequencies sequence: int64 - name: nl_scores dtype: float32 - name: 0_8_snowclones dtype: int64 - name: 0_9_snowclones dtype: int64 - name: 0_8_templates dtype: int64 - name: 0_9_templates dtype: int64 - name: huffman_coding_length dtype: float64 - name: memorization_score dtype: float64 - name: index dtype: int64 - name: loss dtype: float32 - name: prompt_perplexity dtype: float32 - name: generation_perplexity dtype: float32 - name: sequence_perplexity dtype: float32 splits: - name: pile_duped_6.9b num_bytes: 3952510656 num_examples: 5000000 - name: memories_duped_1b num_bytes: 975471445 num_examples: 1256144 - name: pile_deduped_1.4b num_bytes: 3957538226 num_examples: 5000000 - name: memories_duped_6.9b num_bytes: 1639665283 num_examples: 2120976 - name: pile_duped_2.8b num_bytes: 3952510656 num_examples: 5000000 - name: memories_deduped_410m num_bytes: 636147799 num_examples: 811040 - name: memories_deduped_1.4b num_bytes: 819760730 num_examples: 1048104 - name: memories_duped_12b num_bytes: 1840563478 num_examples: 2382328 - name: memories_deduped_6.9b num_bytes: 1311517858 num_examples: 1680296 - name: pile_duped_160m num_bytes: 3955010656 num_examples: 5000000 - name: memories_deduped_1b num_bytes: 808469787 num_examples: 1032872 - name: memories_duped_160m num_bytes: 540094560 num_examples: 689680 - name: pile_deduped_410m num_bytes: 3957538226 num_examples: 5000000 - name: pile_deduped_2.8b num_bytes: 3957538226 num_examples: 5000000 - name: pile_deduped_160m num_bytes: 3957538226 num_examples: 5000000 - name: memories_duped_1.4b num_bytes: 1065534772 num_examples: 1373728 - name: pile_duped_1.4b num_bytes: 3952510656 num_examples: 5000000 - name: memories_duped_70m num_bytes: 365918006 num_examples: 463960 - name: pile_duped_1b num_bytes: 3952510656 num_examples: 5000000 - name: memories_duped_410m num_bytes: 755826030 num_examples: 970344 - name: pile_deduped_6.9b num_bytes: 3957538226 num_examples: 5000000 - name: pile_deduped_1b num_bytes: 3957538226 num_examples: 5000000 - name: pile_duped_410m num_bytes: 3952510656 num_examples: 5000000 - name: memories_deduped_70m num_bytes: 325241847 num_examples: 411448 - name: memories_deduped_2.8b num_bytes: 1058394856 num_examples: 1355216 - name: pile_duped_12b num_bytes: 3952510656 num_examples: 5000000 - name: pile_deduped_12b num_bytes: 3957538226 num_examples: 5000000 - name: memories_duped_2.8b num_bytes: 1296804964 num_examples: 1675080 - name: memories_deduped_160m num_bytes: 457088933 num_examples: 581200 - name: pile_duped_70m num_bytes: 3955010656 num_examples: 5000000 - name: pile_deduped_70m num_bytes: 3957538226 num_examples: 5000000 - name: memories_deduped_12b num_bytes: 1460036167 num_examples: 1871216 download_size: 36995035353 dataset_size: 80101963738 configs: - config_name: default data_files: - split: pile_duped_6.9b path: data/pile_duped_6.9b-* - split: memories_duped_1b path: data/memories_duped_1b-* - split: pile_deduped_1.4b path: data/pile_deduped_1.4b-* - split: memories_duped_6.9b path: data/memories_duped_6.9b-* - split: pile_duped_2.8b path: data/pile_duped_2.8b-* - split: memories_deduped_410m path: data/memories_deduped_410m-* - split: memories_deduped_1.4b path: data/memories_deduped_1.4b-* - split: memories_duped_12b path: data/memories_duped_12b-* - split: memories_deduped_6.9b path: data/memories_deduped_6.9b-* - split: pile_duped_160m path: data/pile_duped_160m-* - split: memories_deduped_1b path: data/memories_deduped_1b-* - split: memories_duped_160m path: data/memories_duped_160m-* - split: pile_deduped_410m path: data/pile_deduped_410m-* - split: pile_deduped_2.8b path: data/pile_deduped_2.8b-* - split: pile_deduped_160m path: data/pile_deduped_160m-* - split: memories_duped_1.4b path: data/memories_duped_1.4b-* - split: pile_duped_1.4b path: data/pile_duped_1.4b-* - split: memories_duped_70m path: data/memories_duped_70m-* - split: pile_duped_1b path: data/pile_duped_1b-* - split: memories_duped_410m path: data/memories_duped_410m-* - split: pile_deduped_6.9b path: data/pile_deduped_6.9b-* - split: pile_deduped_1b path: data/pile_deduped_1b-* - split: pile_duped_410m path: data/pile_duped_410m-* - split: memories_deduped_70m path: data/memories_deduped_70m-* - split: memories_deduped_2.8b path: data/memories_deduped_2.8b-* - split: pile_duped_12b path: data/pile_duped_12b-* - split: pile_deduped_12b path: data/pile_deduped_12b-* - split: memories_duped_2.8b path: data/memories_duped_2.8b-* - split: memories_deduped_160m path: data/memories_deduped_160m-* - split: pile_duped_70m path: data/pile_duped_70m-* - split: pile_deduped_70m path: data/pile_deduped_70m-* - split: memories_deduped_12b path: data/memories_deduped_12b-* ---
提供机构:
usvsnsp
原始信息汇总

数据集特征

数据集包含以下特征:

  • sequence_id: 数据类型为 int64
  • tokens: 数据类型为 int64 的序列
  • text: 数据类型为 string
  • is_incrementing: 数据类型为 bool
  • is_repeating: 数据类型为 bool
  • sequence_duplicates: 数据类型为 int64
  • max_frequency: 数据类型为 int64
  • avg_frequency: 数据类型为 float64
  • min_frequency: 数据类型为 int64
  • median_frequency: 数据类型为 float64
  • p25_frequency: 数据类型为 int64
  • p75_frequency: 数据类型为 int64
  • frequencies: 数据类型为 int64 的序列
  • nl_scores: 数据类型为 float32
  • 0_8_snowclones: 数据类型为 int64
  • 0_9_snowclones: 数据类型为 int64
  • 0_8_templates: 数据类型为 int64
  • 0_9_templates: 数据类型为 int64
  • huffman_coding_length: 数据类型为 float64
  • memorization_score: 数据类型为 float64
  • index: 数据类型为 int64
  • loss: 数据类型为 float32
  • prompt_perplexity: 数据类型为 float32
  • generation_perplexity: 数据类型为 float32
  • sequence_perplexity: 数据类型为 float32

数据集分割

数据集包含以下分割:

  • pile_duped_6.9b: 字节数为 3952510656,示例数为 5000000
  • memories_duped_1b: 字节数为 975471445,示例数为 1256144
  • pile_deduped_1.4b: 字节数为 3957538226,示例数为 5000000
  • memories_duped_6.9b: 字节数为 1639665283,示例数为 2120976
  • pile_duped_2.8b: 字节数为 3952510656,示例数为 5000000
  • memories_deduped_410m: 字节数为 636147799,示例数为 811040
  • memories_deduped_1.4b: 字节数为 819760730,示例数为 1048104
  • memories_duped_12b: 字节数为 1840563478,示例数为 2382328
  • memories_deduped_6.9b: 字节数为 1311517858,示例数为 1680296
  • pile_duped_160m: 字节数为 3955010656,示例数为 5000000
  • memories_deduped_1b: 字节数为 808469787,示例数为 1032872
  • memories_duped_160m: 字节数为 540094560,示例数为 689680
  • pile_deduped_410m: 字节数为 3957538226,示例数为 5000000
  • pile_deduped_2.8b: 字节数为 3957538226,示例数为 5000000
  • pile_deduped_160m: 字节数为 3957538226,示例数为 5000000
  • memories_duped_1.4b: 字节数为 1065534772,示例数为 1373728
  • pile_duped_1.4b: 字节数为 3952510656,示例数为 5000000
  • memories_duped_70m: 字节数为 365918006,示例数为 463960
  • pile_duped_1b: 字节数为 3952510656,示例数为 5000000
  • memories_duped_410m: 字节数为 755826030,示例数为 970344
  • pile_deduped_6.9b: 字节数为 3957538226,示例数为 5000000
  • pile_deduped_1b: 字节数为 3957538226,示例数为 5000000
  • pile_duped_410m: 字节数为 3952510656,示例数为 5000000
  • memories_deduped_70m: 字节数为 325241847,示例数为 411448
  • memories_deduped_2.8b: 字节数为 1058394856,示例数为 1355216
  • pile_duped_12b: 字节数为 3952510656,示例数为 5000000
  • pile_deduped_12b: 字节数为 3957538226,示例数为 5000000
  • memories_duped_2.8b: 字节数为 1296804964,示例数为 1675080
  • memories_deduped_160m: 字节数为 457088933,示例数为 581200
  • pile_duped_70m: 字节数为 3955010656,示例数为 5000000
  • pile_deduped_70m: 字节数为 3957538226,示例数为 5000000
  • memories_deduped_12b: 字节数为 1460036167,示例数为 1871216

数据集大小

  • 下载大小: 36995035353 字节
  • 数据集大小: 80101963738 字节

配置

  • 配置名称: default
  • 数据文件:
    • pile_duped_6.9b: 路径为 data/pile_duped_6.9b-*
    • memories_duped_1b: 路径为 data/memories_duped_1b-*
    • pile_deduped_1.4b: 路径为 data/pile_deduped_1.4b-*
    • memories_duped_6.9b: 路径为 data/memories_duped_6.9b-*
    • pile_duped_2.8b: 路径为 data/pile_duped_2.8b-*
    • memories_deduped_410m: 路径为 data/memories_deduped_410m-*
    • memories_deduped_1.4b: 路径为 data/memories_deduped_1.4b-*
    • memories_duped_12b: 路径为 data/memories_duped_12b-*
    • memories_deduped_6.9b: 路径为 data/memories_deduped_6.9b-*
    • pile_duped_160m: 路径为 data/pile_duped_160m-*
    • memories_deduped_1b: 路径为 data/memories_deduped_1b-*
    • memories_duped_160m: 路径为 data/memories_duped_160m-*
    • pile_deduped_410m: 路径为 data/pile_deduped_410m-*
    • pile_deduped_2.8b: 路径为 data/pile_deduped_2.8b-*
    • pile_deduped_160m: 路径为 data/pile_deduped_160m-*
    • memories_duped_1.4b: 路径为 data/memories_duped_1.4b-*
    • pile_duped_1.4b: 路径为 data/pile_duped_1.4b-*
    • memories_duped_70m: 路径为 data/memories_duped_70m-*
    • pile_duped_1b: 路径为 data/pile_duped_1b-*
    • memories_duped_410m: 路径为 data/memories_duped_410m-*
    • pile_deduped_6.9b: 路径为 data/pile_deduped_6.9b-*
    • pile_deduped_1b: 路径为 data/pile_deduped_1b-*
    • pile_duped_410m: 路径为 data/pile_duped_410m-*
    • memories_deduped_70m: 路径为 data/memories_deduped_70m-*
    • memories_deduped_2.8b: 路径为 data/memories_deduped_2.8b-*
    • pile_duped_12b: 路径为 data/pile_duped_12b-*
    • pile_deduped_12b: 路径为 data/pile_deduped_12b-*
    • memories_duped_2.8b: 路径为 data/memories_duped_2.8b-*
    • memories_deduped_160m: 路径为 data/memories_deduped_160m-*
    • pile_duped_70m: 路径为 data/pile_duped_70m-*
    • pile_deduped_70m: 路径为 data/pile_deduped_70m-*
    • memories_deduped_12b: 路径为 data/memories_deduped_12b-*
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作