Major changes: - New frontend (platform/web/): Vite + React 18 + TypeScript + Tailwind - 4-module navigation: 数据送标 / 模型管理 / 车队管理 / 系统管理 - Data catalog with charts (DMS/ADAS/Lane 3-tab view) - Quality review workflow (标注质检): Good/Fine/Bad scoring with auto-advance - Audit enhancements: batch operations, rejection categories, Feishu notifications - Operation audit log (操作日志) - World model simulation studio (仿真工坊) - Dataset version management with snapshots and diff - ADAS 7-class dataset integration (138K images organized + compressed) - User management with Feishu integration and pagination - CRUD/search/filter on all pages, card layout redesign - PIL-optimized image overlay rendering - Auto-snapshot on build, in_review workflow stage - Removed embedded algorithm code (now in workspace)
66 lines
2.2 KiB
Markdown
66 lines
2.2 KiB
Markdown
# HSAP 数据目录 · 任务划分
|
||
|
||
## 域(domain)
|
||
|
||
| domain | 含义 | 平台任务 ID |
|
||
|--------|------|-------------|
|
||
| **dms** | 舱内驾驶员监控(人脸/疲劳/分心/DAM) | `ddaw`, `addw`, `addw_face`, `dam`(含 0516/0417 两批次) |
|
||
| **forward** | 前向 ADAS · 交通标志 | `forward`(含检测 + 细分类两种子模式) |
|
||
|
||
数据目录页任务下拉按域分组显示;**forward** 选中后需再选 **子模式**:`detect`(粗检测 4 类)或 `classify`(细分类 ~92 类)。
|
||
|
||
## DAM 任务 `dam`(合并原 dam / dam_0417)
|
||
|
||
同一套 **15 类** 驾驶员监控检测(face、眼嘴、眼镜/烟/手机、driver 等),仅采集批次不同:
|
||
|
||
| 子模式 | 原目录 | 说明 |
|
||
|--------|--------|------|
|
||
| **batch_0516** | `dam/`(dam_0516) | 约 5k 图 |
|
||
| **batch_0417** | `dam_0417/` | 约 2.5k 图 |
|
||
|
||
训练 yaml:`dam__batch_0516.yaml`、`dam__batch_0417.yaml`。别名:`dam_0417` → `dam` + `batch_0417`。
|
||
|
||
```bash
|
||
python datasets/dms/scripts/migrate_dam_layout.py \
|
||
--pack-dir /path/to/packs/dms_v1 --dms-root /path/to/datasets/dms
|
||
```
|
||
|
||
## 前向任务 `forward`(合并原 isa / isa_class)
|
||
|
||
| 子模式 | 原任务 ID | 目录(包内) | 说明 |
|
||
|--------|-----------|--------------|------|
|
||
| **detect** | `isa` | `forward/detect/` | YOLO 检测:indicative / prohibitory / warning / vehicle |
|
||
| **classify** | `isa_class` | `forward/classify/` | 文件夹分类,具体牌型 |
|
||
|
||
训练 yaml:`manifests/yaml_active/forward__detect.yaml`、`forward__classify.yaml`
|
||
训练命令示例:
|
||
|
||
```bash
|
||
SUBMODE=detect ./scripts/train.sh forward full
|
||
SUBMODE=classify ./scripts/train.sh forward full
|
||
```
|
||
|
||
入库 inbox:
|
||
|
||
- `datasets/dms/inbox/forward/detect/<batch>`
|
||
- `datasets/dms/inbox/forward/classify/<batch>`
|
||
|
||
## 迁移旧目录
|
||
|
||
若数据仍在 `packs/dms_v1/isa` 与 `isa_class`:
|
||
|
||
```bash
|
||
python datasets/dms/scripts/migrate_forward_layout.py \
|
||
--pack-dir /path/to/packs/dms_v1 \
|
||
--dms-root /path/to/datasets/dms
|
||
```
|
||
|
||
默认创建符号链接,不搬动原数据。完成后执行 `refresh_yaml.py` 并在平台 **数据目录** 点刷新。
|
||
|
||
## 兼容别名
|
||
|
||
脚本与 API 仍接受旧 ID,会自动映射:
|
||
|
||
- `isa` → `forward` + `detect`
|
||
- `isa_class` → `forward` + `classify`
|