datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MultiCloudSRE-Ops
Multi-Cloud SRE Challenge Dataset
多云运维 SRE 故障排查挑战数据集，基于真实的三朵云（阿里云、腾讯云、AWS）电商微服务系统。
Dataset Structure
每条数据包含以下字段：
字段
类型
说明
canary
string
题目唯一标识
tags
array
分类标签 [layer, sub_category, faults..., difficulty, scope]
case
object
完整的 case 数据（包含注入脚本、恢复脚本、故障现象等）
ideal_answer
object
理想答案（包含故障信息、推理过程、验证方法、解决方案等）
rubrics
array
评分标准（每条含 criterion、points、tags）
prompt
string
满分答案 prompt（待收集）
Example
from datasets import load_dataset
ds =… See the full description on the dataset page: https://huggingface.co/datasets/kluoms/MultiCloudSRE-Ops.terraform-multicloud
terraform-multicloud
78,912 natural-language → Terraform (HCL) pairs, drawn from 25,264 public
GitHub repositories across 15 provider families. Every target is human-written
HCL that passes terraform validate. Every row carries its source repository,
licence and file path.
Built by the SASVA AI Model Cognition Labs (MCL) Team and used to train
SASVAAI/qwen38-27b-terraform.
What one row looks like
One complete row from data/train.jsonl, reproduced in full with… See the full description on the dataset page: https://huggingface.co/datasets/SASVAAI/terraform-multicloud.