Commit ·
f72e1f4
0
Parent(s):
Super-squash branch 'main' using huggingface_hub
Browse files- .gitattributes +61 -0
- README.md +163 -0
- dataset_manifest.json +38 -0
- golden_test.jsonl +0 -0
- intent_schema.json +325 -0
- sft_test.jsonl +0 -0
- sft_train.jsonl +3 -0
- sft_valid.jsonl +0 -0
.gitattributes
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.avro filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.lz4 filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.mds filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
# Audio files - uncompressed
|
| 40 |
+
*.pcm filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
*.sam filter=lfs diff=lfs merge=lfs -text
|
| 42 |
+
*.raw filter=lfs diff=lfs merge=lfs -text
|
| 43 |
+
# Audio files - compressed
|
| 44 |
+
*.aac filter=lfs diff=lfs merge=lfs -text
|
| 45 |
+
*.flac filter=lfs diff=lfs merge=lfs -text
|
| 46 |
+
*.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 47 |
+
*.ogg filter=lfs diff=lfs merge=lfs -text
|
| 48 |
+
*.wav filter=lfs diff=lfs merge=lfs -text
|
| 49 |
+
# Image files - uncompressed
|
| 50 |
+
*.bmp filter=lfs diff=lfs merge=lfs -text
|
| 51 |
+
*.gif filter=lfs diff=lfs merge=lfs -text
|
| 52 |
+
*.png filter=lfs diff=lfs merge=lfs -text
|
| 53 |
+
*.tiff filter=lfs diff=lfs merge=lfs -text
|
| 54 |
+
# Image files - compressed
|
| 55 |
+
*.jpg filter=lfs diff=lfs merge=lfs -text
|
| 56 |
+
*.jpeg filter=lfs diff=lfs merge=lfs -text
|
| 57 |
+
*.webp filter=lfs diff=lfs merge=lfs -text
|
| 58 |
+
# Video files - compressed
|
| 59 |
+
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 60 |
+
*.webm filter=lfs diff=lfs merge=lfs -text
|
| 61 |
+
sft_train.jsonl filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,163 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
pretty_name: IntentBench v1
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
license: other
|
| 6 |
+
task_categories:
|
| 7 |
+
- text-generation
|
| 8 |
+
tags:
|
| 9 |
+
- intent-based-networking
|
| 10 |
+
- network-automation
|
| 11 |
+
- network-diagnosis
|
| 12 |
+
- closed-loop-recovery
|
| 13 |
+
- agentic-ai
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
# IntentBench v1
|
| 17 |
+
|
| 18 |
+
## 数据集简介
|
| 19 |
+
|
| 20 |
+
IntentBench v1 是面向意图驱动自愈网络的复合数据集,覆盖从自然语言业务意图到策略配置、状态诊断和闭环恢复的完整链路。数据集用于训练和评测模型的结构化决策、工具编排、证据约束与安全执行能力,而非一般网络问答。
|
| 21 |
+
|
| 22 |
+
数据集包含 **100,000 条记录**,全部完成构建并按统一规范验收。核心场景覆盖校园网/IP-SDN 与 5G/O-RAN;O-RAN 部分仅用于意图与保障评估。
|
| 23 |
+
|
| 24 |
+
## 任务与规模
|
| 25 |
+
|
| 26 |
+
| 子任务 | 独立样例 | 关联记录 | 合计 | 目标输出 |
|
| 27 |
+
| --- | ---: | ---: | ---: | --- |
|
| 28 |
+
| T1 意图识别与规范化 | 20,000 | 5,000 | 25,000 | `StructuredIntent` |
|
| 29 |
+
| T2 策略与配置生成 | 20,000 | 5,000 | 25,000 | `PolicyConfigPlan` |
|
| 30 |
+
| T3 状态评估与诊断 | 20,000 | 5,000 | 25,000 | `Diagnosis` |
|
| 31 |
+
| T4 闭环自愈 | 20,000 | 5,000 | 25,000 | `RecoveryPlan` |
|
| 32 |
+
| **合计** | **80,000** | **20,000** | **100,000** | — |
|
| 33 |
+
|
| 34 |
+
20,000 条关联记录来自 **5,000 个完整闭环场景**,每个场景按 T1~T4 展开为四条记录。其中包含 3,500 个可执行 IP-SDN 场景和 1,500 个 O-RAN 保障评估场景。
|
| 35 |
+
|
| 36 |
+
按构建方式统计,32,500 条记录来自公开资产的清洗与标准化,40,500 条通过规则派生、配置编译、窗口切分或受控改写形成,27,000 条的核心状态与执行反馈由仿真生成。
|
| 37 |
+
|
| 38 |
+
## 数据划分
|
| 39 |
+
|
| 40 |
+
| 数据形态 | Train | Dev | Test | 合计 |
|
| 41 |
+
| --- | ---: | ---: | ---: | ---: |
|
| 42 |
+
| 独立样例 | 64,000 | 8,000 | 8,000 | 80,000 |
|
| 43 |
+
| 关联记录 | 16,000 | 2,000 | 2,000 | 20,000 |
|
| 44 |
+
| 完整场景 | 4,000 | 500 | 500 | 5,000 |
|
| 45 |
+
|
| 46 |
+
划分以 `global_leakage_group` 为最小单位。同一原始记录、内容哈希、模板父族、拓扑、重叠遥测窗口、流量画像、故障事件及关联轨迹必须进入同一 split,避免跨集合泄漏。
|
| 47 |
+
|
| 48 |
+
## 数据格式
|
| 49 |
+
|
| 50 |
+
Canonical Benchmark JSONL 是评测与审计的权威存储,保存任务输入、金标、来源、关联关系、安全约束与 sidecar 引用。数据集另派生四种视图:
|
| 51 |
+
|
| 52 |
+
- `sft_atomic`:四任务的单阶段 `system/user/assistant` 样例;
|
| 53 |
+
- `sft_trajectory`:按 `scenario_id` 关联的四阶段轨迹;
|
| 54 |
+
- `rl_prompts`:移除金标与隐藏检查的强化学习输入;
|
| 55 |
+
- `evaluation`:测试输入、预测 Schema 与内部评分视图。
|
| 56 |
+
|
| 57 |
+
共享顶层字段包括 `schema_version`、`task_id`、`subtask`、`split`、`language`、`domain`、`execution_mode`、`scenario_id`、`trajectory_id`、`global_leakage_group`、`nl_intent`、`network_context`、`current_config`、`telemetry`、`allowed_tools`、`risk_level`、`authorized_scope`、`provenance`、`asset_refs`、`hidden_checks` 及对应的 `gold_*` 字段。
|
| 58 |
+
|
| 59 |
+
关联样例必须具有 `scenario_id` 和 `trajectory_id`;独立样例对应字段为 `null`。必需数组无元素时使用 `[]`,未知状态使用显式 `unknown`,公开测试输入直接移除 `gold_*` 与 `hidden_checks`。
|
| 60 |
+
|
| 61 |
+
来源追踪信息统一放入 `provenance`:`raw_record_id`、`source_record_sha256`、`template_family`、`semantic_frame` 和 `derived_from`。
|
| 62 |
+
|
| 63 |
+
大体积拓扑、配置、遥测、流量画像、工具轨迹及执行证据使用内容寻址 sidecar。Canonical `AssetRef` 格式如下:
|
| 64 |
+
|
| 65 |
+
```json
|
| 66 |
+
{
|
| 67 |
+
"asset_id": "sha256:<64-hex>",
|
| 68 |
+
"uri": "sidecar/sha256/ab/<remaining-hash>",
|
| 69 |
+
"sha256": "<64-hex>",
|
| 70 |
+
"media_type": "application/json",
|
| 71 |
+
"byte_size": 18452,
|
| 72 |
+
"role": "telemetry"
|
| 73 |
+
}
|
| 74 |
+
```
|
| 75 |
+
|
| 76 |
+
同一内容仅保存一次;引用哈希必须与实际字节一致;公开包只携带获准公开且被记录引用的 sidecar。
|
| 77 |
+
|
| 78 |
+
## 任务 Schema 与节选示例
|
| 79 |
+
|
| 80 |
+
以下均为**节选示例,仅展示任务核心字段**。
|
| 81 |
+
|
| 82 |
+
### T1:意图识别与规范化
|
| 83 |
+
|
| 84 |
+
输入为自然语言意图、请求者角色、授权范围及可选网络上下文。输出包含 `intent_type`、`entities`、`constraints`、`priority`、`conflict_status`、`capability_status`、`expected_behavior`、`authorization` 和 `evidence_spans`。
|
| 85 |
+
|
| 86 |
+
```json
|
| 87 |
+
{"subtask":"intent_normalization","nl_intent":"Keep exam latency <= 20 ms and block guest access to 203.0.113.0/24.","gold_structured_intent":{"intent_type":"access_control","constraints":[{"name":"latency","operator":"lte","value":20,"unit":"ms"}],"expected_behavior":"proceed"}}
|
| 88 |
+
```
|
| 89 |
+
|
| 90 |
+
### T2:策略与配置生成
|
| 91 |
+
|
| 92 |
+
输入为结构化意图、拓扑、设备能力、当前配置、授权范围和可用工具。输出包含策略步骤、配置增量、回滚计划、幂等键、验证计划、预期后置条件及拒绝原因。
|
| 93 |
+
|
| 94 |
+
```json
|
| 95 |
+
{"subtask":"policy_config_generation","execution_mode":"executable","allowed_tools":["config_simulation","reachability_check","rollback"],"gold_policy_or_config":{"config_format":"frr","policy_plan":[{"operation":"update","target_type":"acl"}],"expected_postconditions":{"protected_reachability":true}}}
|
| 96 |
+
```
|
| 97 |
+
|
| 98 |
+
### T3:状态评估与诊断
|
| 99 |
+
|
| 100 |
+
输入为结构化意图、部署结果、拓扑、遥测和当前配置。输出包含 `intent_status`、故障、根因、证据锚点、影响范围、下一步建议及缺失观测。
|
| 101 |
+
|
| 102 |
+
```json
|
| 103 |
+
{"subtask":"state_diagnosis","gold_diagnosis":{"intent_status":"violated","root_causes":[{"cause_type":"congestion","entity_ref":"link:r1-r2","confidence":0.96}],"evidence":[{"field":"latency_ms","value":47.2,"threshold":20,"relation":"gt"}],"recommended_next_step":"recover"}}
|
| 104 |
+
```
|
| 105 |
+
|
| 106 |
+
### T4:闭环自愈
|
| 107 |
+
|
| 108 |
+
输入为结构化意图、诊断、拓扑、流量画像、配置、遥测、授权范围和可用工具。输出包含恢复决策、动作序列、目标后置条件、安全不变量、验证窗口、恢复状态、成本估计与证据引用。
|
| 109 |
+
|
| 110 |
+
```json
|
| 111 |
+
{"subtask":"closed_loop_recovery","gold_actions":{"acceptable_action_classes":["reroute","rate_limit"],"forbidden_actions":["disable_acl","remove_protected_route"],"target_postconditions":{"latency_ms_lte":20,"protected_reachability":true}}}
|
| 112 |
+
```
|
| 113 |
+
|
| 114 |
+
## 数据来源
|
| 115 |
+
|
| 116 |
+
| 来源 | 任务与实际数量 | 冻结版本 | 许可证与公开状态 |
|
| 117 |
+
| --- | ---: | --- | --- |
|
| 118 |
+
| [NIT](https://huggingface.co/datasets/Smarneh/NIT) | T1 11,250(45%) | `084554c` | 未声明;`internal_only` |
|
| 119 |
+
| [TMF921 intent-to-config](https://huggingface.co/datasets/nraptisss/TMF921-intent-to-config-research-sota) | T1 8,750;T2 8,750 | `734a972` | Apache-2.0 |
|
| 120 |
+
| [BINS](https://github.com/lijie19740136878/BINS) | T1 2,500(10%) | `f20befc` | 未声明;`internal_only` |
|
| 121 |
+
| [IW-IB-5GNET](https://github.com/jimenaandrade/iw-ib-5gnet) | T1 2,500(10%) | `dceeaec` | GPL-2.0;公开前审查兼容性 |
|
| 122 |
+
| [NetConfEval](https://github.com/RedHatResearch/conext24-NetConfEval) | T2 8,750(35%) | `24edc15` | MIT |
|
| 123 |
+
| [Batfish example networks](https://github.com/batfish/batfish) | T2 7,500(30%) | `efbf7f4` | Apache-2.0 |
|
| 124 |
+
| [SOFI](https://data.mendeley.com/datasets/tc6ysmh5j8/2) | T3 8,750(35%) | v2;`10.17632/tc6ysmh5j8.2` | CC BY 4.0 |
|
| 125 |
+
| [LCORE-D](https://data.mendeley.com/datasets/77sztrg5ks/1) | T3 8,750(35%) | v1;`10.17632/77sztrg5ks.1` | CC BY 4.0 |
|
| 126 |
+
| [RAN KPM OAI](https://zenodo.org/records/17935683) | T3 3,750(15%) | v1;`10.5281/zenodo.17935683` | 权利确认前 `internal_only` |
|
| 127 |
+
| [ColO-RAN](https://github.com/wineslab/colosseum-oran-coloran-dataset) | T3 3,750(15%) | `bd86629` | GPL-3.0;公开前审查兼容性 |
|
| 128 |
+
| [Internet Topology Zoo](https://topology-zoo.org/) | T4 10,000(40%) | `e278b1b` | 再分发许可未确认;`internal_only` |
|
| 129 |
+
| [SNDlib](https://sndlib.put.poznan.pl/networks.overview.action) | T4 8,750(35%) | 冻结下载快照 | 未声明;`internal_only` |
|
| 130 |
+
| [Rocketfuel](https://research.cs.washington.edu/networking/rocketfuel/) | T4 6,250(25%) | `policy-dist.tar.gz` 冻结快照 | 再分发许可未确认;`internal_only` |
|
| 131 |
+
| [Network Digital Twin congestion dataset](https://zenodo.org/records/17255272) | T4 辅助,不单独占配额 | v1;`10.5281/zenodo.17255272` | 权利确认前不进入公开包 |
|
| 132 |
+
|
| 133 |
+
原始资产经字段映射、格式与单位统一、受控派生、配置编译、故障注入或仿真执行后形成标准记录。许可证明确并兼容的内容进入 `public_cleared` 候选;任何关键资产不可公开时,整个 `global_leakage_group` 不进入公开包。
|
| 134 |
+
|
| 135 |
+
## 安全负样本
|
| 136 |
+
|
| 137 |
+
数据集中包含 **20,000 条安全负样本**,已计入10万条总量,四任务各5,000条。风险覆盖:冲突或不安全、需澄清、超范围/OOD、越权工具调用、缺少回滚、破坏可达性、伪造执行结果以及 PRECHECK/POSTCHECK 违规。
|
| 138 |
+
|
| 139 |
+
安全硬门槛优先于局部性能收益。扩大访问面、破坏受保护可达性、越权调用或缺少必要回滚的方案均判为失败。
|
| 140 |
+
|
| 141 |
+
## 质量控制
|
| 142 |
+
|
| 143 |
+
所有记录依次接受 JSON Schema、证据对应、配置解析、可达性、安全约束、轨迹一致性、内容哈希和全局防泄漏检查。配置与闭环场景使用 Batfish、FRR、Kathará 或相应仿真后端验证;缺少工具反馈时不得生成“执行成功”金标。
|
| 144 |
+
|
| 145 |
+
## 推荐用途
|
| 146 |
+
|
| 147 |
+
- 使用64,000条训练集独立样例进行四任务 SFT;
|
| 148 |
+
- 使用4,000个训练集完整场景(16,000条关联记录)进行GRPO或其他可验证强化学习;
|
| 149 |
+
- 使用独立测试集评估单阶段能力,使用关联测试场景评估端到端Agent能力;
|
| 150 |
+
- 用于结构化意图、配置生成、证据化诊断、工具编排、安全恢复和网络智能体评测。
|
| 151 |
+
|
| 152 |
+
## 局限与使用约束
|
| 153 |
+
|
| 154 |
+
- 部分状态和执行反馈来自仿真,不能完全代表生产网络;
|
| 155 |
+
- 数据分布受选定拓扑、设备生态、业务模板和故障类型影响;
|
| 156 |
+
- v1以英文为主,跨语言泛化能力需单独验证;
|
| 157 |
+
- O-RAN样例仅支持保障评估,不支持真实RAN控制;
|
| 158 |
+
- 数据集和模型输出不得用于绕过审批、安全门禁或回滚机制直接修改生产网络;
|
| 159 |
+
- `internal_full` 是完整权威版本,`public_cleared` 仅包含许可证允许公开的完整泄漏组。
|
| 160 |
+
|
| 161 |
+
## 引用
|
| 162 |
+
|
| 163 |
+
如在研究中使用 IntentBench,请引用数据集发布页及 IntentClaw 项目说明。正式引文信息以发布版本中的 `CITATION.cff` 为准。
|
dataset_manifest.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"created_at": "2026-06-03 15:58:13",
|
| 3 |
+
"exact_dedup": 10,
|
| 4 |
+
"golden_train_group_overlap": 0,
|
| 5 |
+
"group_leaks": [],
|
| 6 |
+
"input_paths": [
|
| 7 |
+
"${INTENTCLAW_DATA_ROOT}/generated/manual_schema_examples.jsonl",
|
| 8 |
+
"${INTENTCLAW_DATA_ROOT}/normalized/public_tmf921.jsonl",
|
| 9 |
+
"${INTENTCLAW_DATA_ROOT}/normalized/public_nit.jsonl",
|
| 10 |
+
"${INTENTCLAW_DATA_ROOT}/normalized/public_ibns.jsonl",
|
| 11 |
+
"${INTENTCLAW_DATA_ROOT}/normalized/public_bins.jsonl",
|
| 12 |
+
"${INTENTCLAW_DATA_ROOT}/generated/template_samples.jsonl",
|
| 13 |
+
"${INTENTCLAW_DATA_ROOT}/generated/paraphrase_samples.jsonl",
|
| 14 |
+
"${INTENTCLAW_DATA_ROOT}/generated/safety_negative_samples.jsonl"
|
| 15 |
+
],
|
| 16 |
+
"invalid": 0,
|
| 17 |
+
"loaded": 156335,
|
| 18 |
+
"near_dedup": 0,
|
| 19 |
+
"public_candidate_count": 42805,
|
| 20 |
+
"public_retained_count": 42805,
|
| 21 |
+
"public_retained_target": 35000,
|
| 22 |
+
"schema_pass_rate": 0.9999360347970704,
|
| 23 |
+
"split_counts": {
|
| 24 |
+
"golden": 800,
|
| 25 |
+
"test": 5000,
|
| 26 |
+
"train": 95000,
|
| 27 |
+
"valid": 5001
|
| 28 |
+
},
|
| 29 |
+
"targets": {
|
| 30 |
+
"golden": 800,
|
| 31 |
+
"test": 5000,
|
| 32 |
+
"train": 95000,
|
| 33 |
+
"valid": 5000
|
| 34 |
+
},
|
| 35 |
+
"unused_after_split": 50524,
|
| 36 |
+
"valid_after_clean": 156325,
|
| 37 |
+
"valid_json_rate": 1.0
|
| 38 |
+
}
|
golden_test.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
intent_schema.json
ADDED
|
@@ -0,0 +1,325 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"$id": "https://intentclaw.local/schemas/intent-json-v1.json",
|
| 3 |
+
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
| 4 |
+
"additionalProperties": false,
|
| 5 |
+
"properties": {
|
| 6 |
+
"approval_required": {
|
| 7 |
+
"type": "boolean"
|
| 8 |
+
},
|
| 9 |
+
"audit_required": {
|
| 10 |
+
"type": "boolean"
|
| 11 |
+
},
|
| 12 |
+
"audit_tags": {
|
| 13 |
+
"items": {
|
| 14 |
+
"type": "string"
|
| 15 |
+
},
|
| 16 |
+
"type": "array"
|
| 17 |
+
},
|
| 18 |
+
"constraints": {
|
| 19 |
+
"items": {
|
| 20 |
+
"type": "string"
|
| 21 |
+
},
|
| 22 |
+
"type": "array"
|
| 23 |
+
},
|
| 24 |
+
"entities": {
|
| 25 |
+
"additionalProperties": {
|
| 26 |
+
"type": [
|
| 27 |
+
"string",
|
| 28 |
+
"number",
|
| 29 |
+
"integer",
|
| 30 |
+
"boolean",
|
| 31 |
+
"null"
|
| 32 |
+
]
|
| 33 |
+
},
|
| 34 |
+
"properties": {
|
| 35 |
+
"alert_id": {
|
| 36 |
+
"type": [
|
| 37 |
+
"string",
|
| 38 |
+
"null"
|
| 39 |
+
]
|
| 40 |
+
},
|
| 41 |
+
"application": {
|
| 42 |
+
"type": [
|
| 43 |
+
"string",
|
| 44 |
+
"null"
|
| 45 |
+
]
|
| 46 |
+
},
|
| 47 |
+
"as_number": {
|
| 48 |
+
"minimum": 1,
|
| 49 |
+
"type": [
|
| 50 |
+
"integer",
|
| 51 |
+
"null"
|
| 52 |
+
]
|
| 53 |
+
},
|
| 54 |
+
"bandwidth_mbps_min": {
|
| 55 |
+
"minimum": 0,
|
| 56 |
+
"type": "number"
|
| 57 |
+
},
|
| 58 |
+
"certificate_domain": {
|
| 59 |
+
"type": [
|
| 60 |
+
"string",
|
| 61 |
+
"null"
|
| 62 |
+
]
|
| 63 |
+
},
|
| 64 |
+
"cidr": {
|
| 65 |
+
"type": [
|
| 66 |
+
"string",
|
| 67 |
+
"null"
|
| 68 |
+
]
|
| 69 |
+
},
|
| 70 |
+
"destination": {
|
| 71 |
+
"type": [
|
| 72 |
+
"string",
|
| 73 |
+
"null"
|
| 74 |
+
]
|
| 75 |
+
},
|
| 76 |
+
"destination_cidr": {
|
| 77 |
+
"type": [
|
| 78 |
+
"string",
|
| 79 |
+
"null"
|
| 80 |
+
]
|
| 81 |
+
},
|
| 82 |
+
"destination_zone": {
|
| 83 |
+
"type": [
|
| 84 |
+
"string",
|
| 85 |
+
"null"
|
| 86 |
+
]
|
| 87 |
+
},
|
| 88 |
+
"device": {
|
| 89 |
+
"type": [
|
| 90 |
+
"string",
|
| 91 |
+
"null"
|
| 92 |
+
]
|
| 93 |
+
},
|
| 94 |
+
"event_type": {
|
| 95 |
+
"type": [
|
| 96 |
+
"string",
|
| 97 |
+
"null"
|
| 98 |
+
]
|
| 99 |
+
},
|
| 100 |
+
"interface": {
|
| 101 |
+
"type": [
|
| 102 |
+
"string",
|
| 103 |
+
"null"
|
| 104 |
+
]
|
| 105 |
+
},
|
| 106 |
+
"ip_pool": {
|
| 107 |
+
"type": [
|
| 108 |
+
"string",
|
| 109 |
+
"null"
|
| 110 |
+
]
|
| 111 |
+
},
|
| 112 |
+
"jitter_ms_max": {
|
| 113 |
+
"minimum": 0,
|
| 114 |
+
"type": "number"
|
| 115 |
+
},
|
| 116 |
+
"latency_ms_max": {
|
| 117 |
+
"minimum": 0,
|
| 118 |
+
"type": "number"
|
| 119 |
+
},
|
| 120 |
+
"link_id": {
|
| 121 |
+
"type": [
|
| 122 |
+
"string",
|
| 123 |
+
"null"
|
| 124 |
+
]
|
| 125 |
+
},
|
| 126 |
+
"maintenance_window": {
|
| 127 |
+
"type": [
|
| 128 |
+
"string",
|
| 129 |
+
"null"
|
| 130 |
+
]
|
| 131 |
+
},
|
| 132 |
+
"packet_loss_pct_max": {
|
| 133 |
+
"minimum": 0,
|
| 134 |
+
"type": "number"
|
| 135 |
+
},
|
| 136 |
+
"peer_ip": {
|
| 137 |
+
"type": [
|
| 138 |
+
"string",
|
| 139 |
+
"null"
|
| 140 |
+
]
|
| 141 |
+
},
|
| 142 |
+
"port": {
|
| 143 |
+
"maximum": 65535,
|
| 144 |
+
"minimum": 1,
|
| 145 |
+
"type": "integer"
|
| 146 |
+
},
|
| 147 |
+
"protected_zone": {
|
| 148 |
+
"type": [
|
| 149 |
+
"string",
|
| 150 |
+
"null"
|
| 151 |
+
]
|
| 152 |
+
},
|
| 153 |
+
"protocol": {
|
| 154 |
+
"type": [
|
| 155 |
+
"string",
|
| 156 |
+
"null"
|
| 157 |
+
]
|
| 158 |
+
},
|
| 159 |
+
"service_chain": {
|
| 160 |
+
"type": [
|
| 161 |
+
"string",
|
| 162 |
+
"null"
|
| 163 |
+
]
|
| 164 |
+
},
|
| 165 |
+
"site": {
|
| 166 |
+
"type": [
|
| 167 |
+
"string",
|
| 168 |
+
"null"
|
| 169 |
+
]
|
| 170 |
+
},
|
| 171 |
+
"slice_id": {
|
| 172 |
+
"type": [
|
| 173 |
+
"string",
|
| 174 |
+
"null"
|
| 175 |
+
]
|
| 176 |
+
},
|
| 177 |
+
"source": {
|
| 178 |
+
"type": [
|
| 179 |
+
"string",
|
| 180 |
+
"null"
|
| 181 |
+
]
|
| 182 |
+
},
|
| 183 |
+
"source_cidr": {
|
| 184 |
+
"type": [
|
| 185 |
+
"string",
|
| 186 |
+
"null"
|
| 187 |
+
]
|
| 188 |
+
},
|
| 189 |
+
"source_zone": {
|
| 190 |
+
"type": [
|
| 191 |
+
"string",
|
| 192 |
+
"null"
|
| 193 |
+
]
|
| 194 |
+
},
|
| 195 |
+
"target_ip": {
|
| 196 |
+
"type": [
|
| 197 |
+
"string",
|
| 198 |
+
"null"
|
| 199 |
+
]
|
| 200 |
+
},
|
| 201 |
+
"tenant": {
|
| 202 |
+
"type": [
|
| 203 |
+
"string",
|
| 204 |
+
"null"
|
| 205 |
+
]
|
| 206 |
+
},
|
| 207 |
+
"threat_type": {
|
| 208 |
+
"type": [
|
| 209 |
+
"string",
|
| 210 |
+
"null"
|
| 211 |
+
]
|
| 212 |
+
},
|
| 213 |
+
"vlan_id": {
|
| 214 |
+
"maximum": 4094,
|
| 215 |
+
"minimum": 1,
|
| 216 |
+
"type": "integer"
|
| 217 |
+
},
|
| 218 |
+
"vrf": {
|
| 219 |
+
"type": [
|
| 220 |
+
"string",
|
| 221 |
+
"null"
|
| 222 |
+
]
|
| 223 |
+
}
|
| 224 |
+
},
|
| 225 |
+
"type": "object"
|
| 226 |
+
},
|
| 227 |
+
"expected_output": {
|
| 228 |
+
"type": "string"
|
| 229 |
+
},
|
| 230 |
+
"intent_type": {
|
| 231 |
+
"enum": [
|
| 232 |
+
"qos_assurance",
|
| 233 |
+
"access_control",
|
| 234 |
+
"traffic_steering",
|
| 235 |
+
"network_slice",
|
| 236 |
+
"service_chain",
|
| 237 |
+
"incident_diagnosis",
|
| 238 |
+
"capacity_scaling",
|
| 239 |
+
"observability",
|
| 240 |
+
"configuration_change",
|
| 241 |
+
"security_compliance",
|
| 242 |
+
"rollback_or_recovery",
|
| 243 |
+
"clarification_required",
|
| 244 |
+
"conflict_or_unsafe",
|
| 245 |
+
"out_of_scope"
|
| 246 |
+
],
|
| 247 |
+
"type": "string"
|
| 248 |
+
},
|
| 249 |
+
"missing_fields": {
|
| 250 |
+
"items": {
|
| 251 |
+
"type": "string"
|
| 252 |
+
},
|
| 253 |
+
"type": "array"
|
| 254 |
+
},
|
| 255 |
+
"precheck_required": {
|
| 256 |
+
"type": "boolean"
|
| 257 |
+
},
|
| 258 |
+
"recommended_tools": {
|
| 259 |
+
"items": {
|
| 260 |
+
"enum": [
|
| 261 |
+
"nl-parser",
|
| 262 |
+
"diff-checker",
|
| 263 |
+
"conflict-analyzer",
|
| 264 |
+
"sandbox-simulator",
|
| 265 |
+
"compliance-audit",
|
| 266 |
+
"netconf-pusher",
|
| 267 |
+
"ssh-commander",
|
| 268 |
+
"vlan-manager",
|
| 269 |
+
"qos-optimizer",
|
| 270 |
+
"acl-guard",
|
| 271 |
+
"bgp-neighbor-tool",
|
| 272 |
+
"zero-touch-provisioner",
|
| 273 |
+
"traffic-watcher",
|
| 274 |
+
"path-tracer",
|
| 275 |
+
"optical-doctor",
|
| 276 |
+
"flow-analyzer",
|
| 277 |
+
"health-scorer",
|
| 278 |
+
"root-cause-engine",
|
| 279 |
+
"auto-rerouter",
|
| 280 |
+
"port-recov-tool",
|
| 281 |
+
"congestion-solver",
|
| 282 |
+
"rollback-agent",
|
| 283 |
+
"psirt-scanner",
|
| 284 |
+
"cloud-bridge",
|
| 285 |
+
"ipam-master",
|
| 286 |
+
"ddos-shield",
|
| 287 |
+
"power-saver",
|
| 288 |
+
"cert-manager",
|
| 289 |
+
"forecast-predictor",
|
| 290 |
+
"wiki-rag-search"
|
| 291 |
+
],
|
| 292 |
+
"type": "string"
|
| 293 |
+
},
|
| 294 |
+
"minItems": 1,
|
| 295 |
+
"type": "array",
|
| 296 |
+
"uniqueItems": true
|
| 297 |
+
},
|
| 298 |
+
"requires_confirmation": {
|
| 299 |
+
"type": "boolean"
|
| 300 |
+
},
|
| 301 |
+
"risk_level": {
|
| 302 |
+
"enum": [
|
| 303 |
+
"low",
|
| 304 |
+
"medium",
|
| 305 |
+
"high",
|
| 306 |
+
"critical"
|
| 307 |
+
],
|
| 308 |
+
"type": "string"
|
| 309 |
+
},
|
| 310 |
+
"rollback_required": {
|
| 311 |
+
"type": "boolean"
|
| 312 |
+
}
|
| 313 |
+
},
|
| 314 |
+
"required": [
|
| 315 |
+
"intent_type",
|
| 316 |
+
"entities",
|
| 317 |
+
"constraints",
|
| 318 |
+
"missing_fields",
|
| 319 |
+
"risk_level",
|
| 320 |
+
"recommended_tools",
|
| 321 |
+
"requires_confirmation"
|
| 322 |
+
],
|
| 323 |
+
"title": "IntentClaw IntentJson SFT Output",
|
| 324 |
+
"type": "object"
|
| 325 |
+
}
|
sft_test.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
sft_train.jsonl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2f497bf19d5681e9ecc27aa8d4df649745070b6f2f81e3745abb70d0987b0c4
|
| 3 |
+
size 125278019
|
sft_valid.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|