Spaces:
Sleeping
Sleeping
Commit History
-
Add API key input to UI and run pipeline in thread to avoid deadlock
15be885 -
Updated README with metadata
d4aa497 -
Resolve README.md conflict: keep original README for project
d900c13 -
Phase 4: Add HuggingFace Spaces deployment (Groq fallback, static UI, /run_pipeline endpoint, Dockerfile non-root user)
569851b -
initial commit
Verified signatureec4c909
-
Remove Done: true/false from executor output
8ba6403 -
Phase 2: Add comprehensive grader test suite
- NEW: test_graders.py with 13 scenarios - Tests all 3 graders: CrashGrader, CascadeGrader, SilentDegradeGrader - Validates correct answers score high (0.9-1.0) - Validates wrong answers are properly penalized (0.0-0.5) - All 13/13 tests passinge6bce11 -
Phase 1: Add silent degradation grader (P2 severity)
- NEW: silent_degrade_grader.py for silent degradation task - P2 severity (not P1), payment-db root cause, flush-cache/kill-query remediation - Differentiates slow performance from crashes - Strict validation: -0.15 penalty for P1 overcall, -0.10 for service confusion4cc3c1e -
Phase 1: Fixed graders with strict validation rules
- crash_grader: P1, payment-service, restart validation - cascade_grader: P1, user-db (not payment-service), ordering penalty - crash_grader & cascade_grader: Replace Unicode ≤ with ASCII <= - Ensures honest scoring - no partial credit for wrong answers - All 3 graders now strictly enforce correct vs incorrect answers97662da -
Phase 3: Docker, demo, final cleanup — submission ready
b4feb46 -
Phase 2: Multi-agent pipeline — Planner, Executor, Summarizer, LangGraph
f61dd2d -
Phase 1: Repo setup, environment port, AMD LLM client
4c6fee2 -
Phase 1: Repo setup, environment port, AMD LLM client
658961d
-
Update README.md
Unverified signature28d8d4c -
Phase 1: repo setup and environment port
c41eb02 -
Initial commit
Unverified signature6ce1392