vwxyzjn/ultrachat_200k_filtered_1707920039
收藏Hugging Face2024-02-14 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/vwxyzjn/ultrachat_200k_filtered_1707920039
下载链接
链接失效反馈官方服务:
资源简介:
---
dataset_info:
features:
- name: prompt
dtype: string
- name: prompt_id
dtype: string
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: query_reference_response
list:
- name: content
dtype: string
- name: role
dtype: string
- name: query_reference_response_token
sequence: int64
- name: query_reference_response_token_len
dtype: int64
- name: query
list:
- name: content
dtype: string
- name: role
dtype: string
- name: query_token
sequence: int64
- name: query_token_len
dtype: int64
- name: reference_response
struct:
- name: content
dtype: string
- name: role
dtype: string
- name: reference_response_token
sequence: int64
- name: reference_response_token_len
dtype: int64
splits:
- name: test_gen
num_bytes: 30484069
num_examples: 1000
- name: test_sft
num_bytes: 39592502
num_examples: 1000
- name: train_gen
num_bytes: 29613744
num_examples: 1000
- name: train_sft
num_bytes: 39521233
num_examples: 1000
download_size: 50859072
dataset_size: 139211548
---
# Args
```python
{'base_model': 'mistralai/Mistral-7B-v0.1',
'check_length_correctness': True,
'debug': True,
'hf_entity': 'vwxyzjn',
'params': TaskQueryHParams(length=3000,
format_str='SUBREDDIT: r/{subreddit}\n'
'\n'
'TITLE: {title}\n'
'\n'
'POST: {post}\n'
'\n'
'TL;DR:',
truncate_field='post',
truncate_text='\n',
padding='pad_token',
pad_token=[32000],
pad_side='left',
max_sft_response_length=1500,
max_sft_query_response_length=4500,
max_rm_response_length=169,
max_rm_query_response_length=638),
'push_to_hub': True}
```
This dataset is primarily designed for natural language processing tasks, featuring multiple attributes such as prompt, prompt_id, messages, etc., each with specific data types. The dataset is divided into several subsets, including test_gen, test_sft, train_gen, and train_sft, each with specified byte sizes and example counts. The datasets download size is 50859072 bytes, and its total size is 139211548 bytes.
提供机构:
vwxyzjn
原始信息汇总
数据集信息
特征
- prompt: 字符串类型
- prompt_id: 字符串类型
- messages: 列表类型,包含以下字段:
- content: 字符串类型
- role: 字符串类型
- query_reference_response: 列表类型,包含以下字段:
- content: 字符串类型
- role: 字符串类型
- query_reference_response_token: 整数序列类型
- query_reference_response_token_len: 64位整数类型
- query: 列表类型,包含以下字段:
- content: 字符串类型
- role: 字符串类型
- query_token: 整数序列类型
- query_token_len: 64位整数类型
- reference_response: 结构类型,包含以下字段:
- content: 字符串类型
- role: 字符串类型
- reference_response_token: 整数序列类型
- reference_response_token_len: 64位整数类型
数据分割
- test_gen: 30484069字节,1000个样本
- test_sft: 39592502字节,1000个样本
- train_gen: 29613744字节,1000个样本
- train_sft: 39521233字节,1000个样本
数据集大小
- 下载大小: 50859072字节
- 数据集大小: 139211548字节



