pyr0ball
e2658f743f
Switches to OrchestratedScheduler in cloud mode so concurrent recipe_llm jobs fan out across all registered cf-orch GPU nodes instead of serializing on one. Under load this eliminates poll timeouts from queue backup. USE_ORCH_SCHEDULER env var gives explicit control independent of CLOUD_MODE: unset follow CLOUD_MODE (cloud=orch, local=local) true OrchestratedScheduler always (e.g. multi-GPU local rig) false LocalScheduler always (e.g. cloud single-GPU dev instance) ImportError fallback: if circuitforge_orch is not installed and orch is requested, logs a warning and falls back to LocalScheduler gracefully. |
||
|---|---|---|
| .. | ||
| api | ||
| core | ||
| db | ||
| models | ||
| services | ||
| staples | ||
| static | ||
| styles | ||
| tasks | ||
| utils | ||
| __init__.py | ||
| cloud_session.py | ||
| main.py | ||
| tiers.py | ||