five

jedidiahkkh/mnli-stats

收藏
Hugging Face2024-05-15 更新2024-06-12 收录
下载链接:
https://hf-mirror.com/datasets/jedidiahkkh/mnli-stats
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集包含多个配置,每个配置具有特定的特征和分割。特征包括epoch、uid、logits、loss以及多种梯度范数和一个标签。每个配置都有不同的epoch作为分割,并指定了字节数和示例数。每个配置的数据文件也列出,将分割映射到各自的路径。

This dataset includes multiple configurations, each with specific features and splits. Features include epoch, uid, logits, loss, and various gradient norms, along with a label. Each configuration has different epochs as splits, with specified numbers of bytes and examples. The data files for each configuration are also listed, mapping splits to their respective paths.
提供机构:
jedidiahkkh
原始信息汇总

数据集概述

数据集配置1

  • config_name: jedidiahkkh__bert-base-uncased_mnli_2f0d7d9525
  • features:
    • epoch: int32
    • uid: int64
    • logits: float64 (sequence)
    • loss: float64
    • grad_1norm: float64
    • grad_2norm: float64
    • grad_infnorm: float64
    • label: int32
  • splits:
    • epoch10: 29712000 bytes, 384000 examples
    • epoch2: 29712000 bytes, 384000 examples
    • epoch5: 29712000 bytes, 384000 examples
    • epoch9: 29712000 bytes, 384000 examples
    • epoch6: 29712000 bytes, 384000 examples
    • epoch3: 29712000 bytes, 384000 examples
    • epoch0: 29712000 bytes, 384000 examples
    • epoch8: 29712000 bytes, 384000 examples
    • epoch7: 29712000 bytes, 384000 examples
    • epoch4: 29712000 bytes, 384000 examples
    • epoch1: 29712000 bytes, 384000 examples
  • download_size: 143859099 bytes
  • dataset_size: 326832000 bytes

数据集配置2

  • config_name: jedidiahkkh__bert-tiny_mnli_204ea4ae32
  • features:
    • epoch: int32
    • uid: int64
    • logits: float64 (sequence)
    • loss: float64
    • grad_1norm: float64
    • grad_2norm: float64
    • grad_infnorm: float64
    • label: int32
  • splits:
    • epoch0: 29712000 bytes, 384000 examples
    • epoch5: 29712000 bytes, 384000 examples
    • epoch6: 29712000 bytes, 384000 examples
    • epoch3: 29712000 bytes, 384000 examples
    • epoch1: 29712000 bytes, 384000 examples
    • epoch2: 29712000 bytes, 384000 examples
    • epoch4: 29712000 bytes, 384000 examples
  • download_size: 95897010 bytes
  • dataset_size: 207984000 bytes

数据集配置3

  • config_name: jedidiahkkh__bert-tiny_mnli_38e115c028
  • features:
    • epoch: int32
    • uid: int64
    • logits: float64 (sequence)
    • loss: float64
    • grad_1norm: float64
    • grad_2norm: float64
    • grad_infnorm: float64
    • label: int32
  • splits:
    • epoch5: 29712000 bytes, 384000 examples
    • epoch1: 29712000 bytes, 384000 examples
    • epoch4: 29712000 bytes, 384000 examples
    • epoch0: 29712000 bytes, 384000 examples
    • epoch6: 29712000 bytes, 384000 examples
    • epoch3: 29712000 bytes, 384000 examples
    • epoch2: 29712000 bytes, 384000 examples
  • download_size: 96172433 bytes
  • dataset_size: 207984000 bytes

数据集配置4

  • config_name: jedidiahkkh__bert-tiny_mnli_ee6146342c
  • features:
    • epoch: int32
    • uid: int64
    • logits: float64 (sequence)
    • loss: float64
    • grad_1norm: float64
    • grad_2norm: float64
    • grad_infnorm: float64
    • label: int32
  • splits:
    • epoch2: 29712000 bytes, 384000 examples
    • epoch10: 29712000 bytes, 384000 examples
    • epoch5: 29712000 bytes, 384000 examples
    • epoch9: 29712000 bytes, 384000 examples
    • epoch4: 29712000 bytes, 384000 examples
    • epoch0: 29712000 bytes, 384000 examples
    • epoch7: 29712000 bytes, 384000 examples
    • epoch8: 29712000 bytes, 384000 examples
    • epoch1: 29712000 bytes, 384000 examples
    • epoch3: 29712000 bytes, 384000 examples
    • epoch6: 29712000 bytes, 384000 examples
  • download_size: 106859198 bytes
  • dataset_size: 326832000 bytes

数据集配置5

  • config_name: jedidiahkkh__roberta-base_mnli_fbca1df448
  • features:
    • epoch: int32
    • uid: int64
    • logits: float64 (sequence)
    • loss: float64
    • grad_1norm: float64
    • grad_2norm: float64
    • grad_infnorm: float64
    • label: int32
  • splits:
    • epoch9: 29712000 bytes, 384000 examples
    • epoch5: 29712000 bytes, 384000 examples
    • epoch4: 29712000 bytes, 384000 examples
    • epoch0: 29712000 bytes, 384000 examples
    • epoch2: 29712000 bytes, 384000 examples
    • epoch10: 29712000 bytes, 384000 examples
    • epoch6: 29712000 bytes, 384000 examples
    • epoch3: 29712000 bytes, 384000 examples
    • epoch8: 29712000 bytes, 384000 examples
    • epoch7: 29712000 bytes, 384000 examples
    • epoch1: 29712000 bytes, 384000 examples
  • download_size: 148640392 bytes
  • dataset_size: 326832000 bytes
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