feat(harness): add Action Recognition training experiment

- PerCategoryPoolStrategy: filter test pool by task_types
- RunConfig: add run_holdout_eval toggle (default true)
- load_config: fix YAML task_types list-to-tuple conversion
- Runner: conditionally skip _holdout_four_way when disabled
- CLI: add --no-run-holdout-eval flag
- New config/train_action_recognition.yaml (3 epochs, per_category)
- New scripts/train_action_recognition.sh (baseline + seed + train)
This commit is contained in:
2026-07-14 00:58:54 -04:00
parent 37d4519905
commit dec7346da3
12 changed files with 1423 additions and 52 deletions
+8
View File
@@ -65,3 +65,11 @@
- [2026-07-12 02:38 UTC] 新增 plan: 出题管线 v2 实现计划 (plan:2026-07-11-question-gen-v2)
- [2026-07-12 02:57 UTC] 新增边: plan:2026-07-11-question-gen-v2 --implements--> design:2026-07-11-question-gen-v2
- [2026-07-12 02:57 UTC] 重建索引: 53 篇页面
- [2026-07-13 02:05 UTC] 新增 design: Per-Category Pool Strategy 设计 (design:per-category-pool-strategy)
- [2026-07-13 02:06 UTC] 重建索引: 55 篇页面
- [2026-07-13 02:26 UTC] 新增 plan: Per-Category Pool Strategy 实现计划 (plan:per-category-pool-strategy)
- [2026-07-13 02:26 UTC] 新增边: plan:per-category-pool-strategy --implements--> design:per-category-pool-strategy
- [2026-07-13 02:27 UTC] 重建索引: 57 篇页面
- [2026-07-14 04:50 UTC] 新增 plan: Action Recognition 单题型首次训练实验计划 (plan:action-recognition-training)
- [2026-07-14 04:50 UTC] 新增边: plan:action-recognition-training --implements--> design:action-recognition-training
- [2026-07-14 04:50 UTC] 重建索引: 60 篇页面