Saguaro-1.0-2b (Proactive Sensor Agent)
This is a frontier research model based on Gemma 3n-E2B, finetuned using LoRA for Proactive Edge Monitoring.
Usage
This model expects a specific input format:
Output
The model will generate:
- REASONING: Analysis of signal vs noise.
- ACTION: A tool call or 'WAIT'.
- SUMMARY: A condensed state to be passed to the next history turn.
Technical Details
- Architecture: Gemma 3n E2B
- Format: JAX/Orbax Checkpoint
- Training: 1800 Steps, Batch Size 4, LoRA Rank 8
