Spaces:
Paused
Paused
Download sensorflow/agentic/agents/__init__.py from whassan/SensorflowStudio: direct link, hf CLI and curl.
- Browser
- Download file 1.32 kB
-
https://huggingface.co/spaces/whassan/SensorflowStudio/resolve/main/sensorflow/agentic/agents/__init__.py
- Command line
-
hf download hf://spaces/whassan/SensorflowStudio/sensorflow/agentic/agents/__init__.py
-
curl -L -o __init__.py https://huggingface.co/spaces/whassan/SensorflowStudio/resolve/main/sensorflow/agentic/agents/__init__.py
1.32 kB
| """The eight advisory agents. | |
| Each agent shares the BaseAgent contract (typed input/output, confidence, | |
| failure handling, escalation rules, human-review triggers) and follows the | |
| platform's copilot pattern: an OPTIONAL local LLM (Ollama) may add a | |
| natural-language rationale, but every agent has a DETERMINISTIC RULE-BASED | |
| core so the whole system runs and tests without any model. Agents are | |
| structurally advisory (AgentResult.authority == "ADVISORY_ONLY"); the | |
| deterministic policy engine in policy.py is the only launch authority. | |
| """ | |
| from sensorflow.agentic.agents.base import BaseAgent # noqa: F401 | |
| from sensorflow.agentic.agents.failure_detection import FailureDetectionAgent # noqa: F401 | |
| from sensorflow.agentic.agents.vlm_scene import VLMSceneAnalysisAgent # noqa: F401 | |
| from sensorflow.agentic.agents.fusion_verification import SensorFusionVerificationAgent # noqa: F401 | |
| from sensorflow.agentic.agents.scenario_mining import ScenarioMiningAgent # noqa: F401 | |
| from sensorflow.agentic.agents.statistical import StatisticalRegressionAgent # noqa: F401 | |
| from sensorflow.agentic.agents.safety_impact import SafetyImpactAgent # noqa: F401 | |
| from sensorflow.agentic.agents.launch_decision import LaunchDecisionAgent # noqa: F401 | |
| from sensorflow.agentic.agents.flywheel_agent import EvalFlywheelAgent # noqa: F401 | |