metadata

dataset_info:
  dataset_size: 1430500
  download_size: 483276
  features:
    - name: input_ids
      sequence: int32
    - name: attention_mask
      sequence: int8
    - name: labels
      sequence: int64
    - dtype: string
      name: preview
  splits:
    - name: train
      num_bytes: 1430500
      num_examples: 400

zh-tw-llm-dev-sample-ta8k-d40d11-only_embeddings-tr__alp-cd978d-c2048

This dataset is a part of the zh-tw-llm-dev project.

Tokenizer: zh-tw-llm-dev-tokenizer-a8k-d40d11
Built with: translations, wikipedia, alpaca
Rows: 400
Max length: 2048

Full config:

{"build_with": ["translations", "wikipedia", "alpaca"], "preview_length": 256, "translations_settings": {"source_dataset": "zetavg/coct-en-zh-tw-translations-twp-300k", "lang_1_key": "en", "lang_2_key": "ch", "templates": ["English: {lang_1}\nChinese: {lang_2}", "Chinese: {lang_2}\nEnglish: {lang_1}"], "rows_limit": 100}, "wikipedia_settings": {"source_dataset": "zetavg/zh-tw-wikipedia", "exclude": [{"match": "小行星", "in": "markdown", "in_range": [0, 40]}, {"match": "是中華人民共和國", "in": "markdown", "in_range": [0, 80]}], "rows_limit": 100}, "alpaca_settings": {"source_dataset": "zetavg/traditional-chinese-alpaca-en-align", "template": "short", "train_on_inputs": false, "rows_limit": 100}}