Fetching metadata from the HF Docker repository...
-
fix: evaluate() uses trained agents, grader uses agent instances in eval loop
-
Fix all audit issues: conflict schema, curriculum wiring, density_factor, eval path, dead files removed
-
feat: implement multi-agent interaction, negotiation, and evaluation framework
-
feat: implement task grading module and update inference logging to match strict output format
-
feat: implement episode grader and update inference scoring to enforce strict (0.01, 0.99) bounds
-
fix: min_reward = steps * -2.0 NOT steps * num_sensors * -2.0 — DO NOT REVERT
-
error fixed
-
grader error fix
-
grader fixed
-
grader fix
# Conflicts:
# tasks/grader.py -
grader fix
-
fix: grader early return also strictly within (0, 1)
-
fix: clamp grader score to strictly open interval (0, 1)
-
fix: correct import in tasks/__init__.py
-
submission ready
-
improve: better prompt + fixed grader normalization, score 0.77
-
improve: better prompt + fixed grader normalization, score 0.77
-
Round 1 submission — SentinelEnv OpenEnv
