Add pipeline/rl/README.md
Browse files- pipeline/rl/README.md +1 -9
pipeline/rl/README.md
CHANGED
|
@@ -1,11 +1,3 @@
|
|
| 1 |
# RL (GRPO/PPO) sketch for tables
|
| 2 |
|
| 3 |
-
|
| 4 |
-
- JSON validity
|
| 5 |
-
- Exact field order
|
| 6 |
-
- Type checks per column
|
| 7 |
-
- Cell‑level F1 / edit distance
|
| 8 |
-
- Row count / primary‑key constraints
|
| 9 |
-
|
| 10 |
-
Loop: rollouts → score → GRPO step → checkpoint → eval.
|
| 11 |
-
Safety: reward clipping, degenerate‑mode detection, strict validators.
|
|
|
|
| 1 |
# RL (GRPO/PPO) sketch for tables
|
| 2 |
|
| 3 |
+
Reward terms to consider: JSON validity, exact key order, type checks, row count.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|