distilabel-internal-testing/pipe_nothing_test

该数据集包含一个pipeline.yaml文件，可用于在distilabel中重现生成数据集的流程。数据集的结构包括两个字段：instruction（指令）和response（响应）。数据集只有一个配置，名为default，可以通过Hugging Face的datasets库加载。

This dataset contains a `pipeline.yaml` which can be used to reproduce the pipeline that generated it in distilabel. The dataset structure includes two fields: instruction and response. There is only one configuration named `default`, and it can be loaded using the Hugging Face datasets library.

原始信息汇总

Dataset Card for pipe_nothing_test

Dataset Summary

• Size Categories: n<1K
• Features:
  • instruction: string
  • response: string
• Splits:
  • train:
    • num_bytes: 64
    • num_examples: 2
• Download Size: 1320
• Dataset Size: 64
• Configs:
  • default:
    • data_files:
      • split: train
      • path: data/train-*
• Tags:
  • synthetic
  • distilabel
  • rlaif

Dataset Structure

• Configuration: default

  • Example: json { "instruction": "Tell me a joke.", "response": "unit test" }

  • Loading: python from datasets import load_dataset

    ds = load_dataset("distilabel-internal-testing/pipe_nothing_test", "default")

    or simply: python from datasets import load_dataset

    ds = load_dataset("distilabel-internal-testing/pipe_nothing_test")