Spaces:
Sleeping
Sleeping
Commit History
-
Fix score range and [END] format for Phase 2 validation
Two issues: 1. Grader could return exactly 0.0 or 1.0; Phase 2 requires strictly (0, 1). Clamped compute_score output to max(0.01, min(0.99, raw)). 2. [END] stdout line was missing required score= field. Sample format: [END] success=... steps=... score=... rewards=... Added score tracking in _agent_loop and updated emit_end signature.07ca292
-
Fix inference Task4 hint, stale docstring, Dockerfile port comment
- inference.py: replace broken CREATE-INSERT-DROP-RENAME hint for Task 4 with correct ALTER TABLE RENAME COLUMN instructions; DROP TABLE on inventory/categories is an instant -1.0 safety violation - server/environment.py: update reset() docstring from seed%3/tasks 1-3 to seed%5/tasks 1-5 - Dockerfile: fix comment example port from 8000 to 7860c979cbb -
Add tasks 4-5, fix metadata, fix port, expand tests to 94
5e2a88b
-
Initial commit
822ca51