Weni/chunks_validation_1.0.0
收藏Hugging Face2024-08-02 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/Weni/chunks_validation_1.0.0
下载链接
链接失效反馈官方服务：
资源简介：
---
language:
- pt
pretty_name: "Chunks Validation 1.0.0"
task_categories:
- question-answering
tags:
- RAG
- cohere
---
```json
{
"id": "chunks_validation_1.0.0",
"name": "Chunks Validation 1.0.0",
"description": "Chunks validation dataset to validate if RAG is extracting the correct small chunks from the source large chunks.",
"task_categories": [
"question-answering"
],
"languages": [
"pt"
],
"dataset": "chunks_validation_1.0.0",
"features": {
"id": "Value(dtype='int64', id=None)",
"base": "Value(dtype='string', id=None)",
"query": "Value(dtype='string', id=None)",
"resposta": "Value(dtype='string', id=None)",
"correct_small_chunks": "Sequence(feature=Value(dtype='string', id=None), length=-1, id=None)",
"source_large_chunks": "Sequence(feature=Value(dtype='string', id=None), length=-1, id=None)",
"small_chunks_with_scores": "[{'content': Value(dtype='string', id=None), 'score': Value(dtype='float64', id=None)}]",
"big_chunks_with_scores": "[{'content': Value(dtype='string', id=None), 'score': Value(dtype='float64', id=None)}]",
"mean_small_chunk_score": "Value(dtype='float64', id=None)",
"min_small_chunk_score": "Value(dtype='float64', id=None)",
"max_small_chunk_score": "Value(dtype='float64', id=None)",
"total_small_chunks": "Value(dtype='int64', id=None)",
"total_correct_small_chunks": "Value(dtype='int64', id=None)",
"cohere_chunks": "[{'chunk': Value(dtype='string', id=None), 'score': Value(dtype='float64', id=None)}]",
"mean_cohere_score": "Value(dtype='float64', id=None)",
"total_cohere_chunks": "Value(dtype='int64', id=None)",
"min_cohere_score": "Value(dtype='float64', id=None)",
"max_cohere_score": "Value(dtype='float64', id=None)"
},
"config": {
"version": "1.0.0",
"split": {
"test": {
"size": 20
}
}
}
}
```
提供机构：
Weni
