Atropos —
Facility-Wide Energy Forecaster
Atropos forecasts whole-facility energy consumption 24 hours ahead for demand charge management and utility rate optimization. Named for the Greek fate who cuts the thread of life — here, cutting energy waste. Deployed alongside Apollo on diagnostic appliances where demand charge management is contractually significant.
Abstract
Background
Atropos forecasts whole-facility energy consumption 24 hours ahead for demand charge management and utility rate optimization. Named for the Greek fate who cuts the thread of life — here, cutting energy waste. Deployed alongside Apollo on diagnostic appliances where demand charge management is contractually significant.
Approach
The model was trained in AxonML (a pure-Rust deep learning framework) and compiled through the Hailo Dataflow Compiler (DFC 3.33.1) targeting Hailo-8 silicon. Post-training INT8 quantization was applied during the DFC compilation pass with production telemetry calibration data. The resulting Hailo Executable Format (HEF) binary executes on Hailo’s fixed-function dataflow architecture with deterministic latency and zero framework overhead at the edge.
Results
On production hardware (Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289)), Atropos achieves 7,026 FPS (hw_only) with 0.178 ms hardware latency at 1.23 W average power draw.
Conclusion
Atropos is production-ready as a single HEF binary deployed to edge devices with no external dependencies beyond the HailoRT vendor runtime. The model meets real-time latency requirements for its target hvac analytics application.
| Model | Atropos |
| Domain | HVAC Analytics |
| Architecture | Facility-Wide Energy Forecaster |
| Target silicon | Hailo-8 |
| Measured on | Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289) |
| DFC compiler | 3.33.1 |
| Framework | AxonML v0.6 (pure-Rust, CUDA + CPU backends) |
| Author | Andrew Jewell Sr. · ORCID 0009-0005-2158-7060 |
| Organization | AutomataNexus LLC · Fort Wayne, Indiana |
Executive overview
Atropos forecasts whole-facility energy consumption 24 hours ahead for demand charge management and utility rate optimization. Named for the Greek fate who cuts the thread of life — here, cutting energy waste. Deployed alongside Apollo on diagnostic appliances where demand charge management is contractually significant.
Network I/O
Input: 22-channel combined telemetry + weather + schedule. Output: 24-hour energy demand forecast.
Architecture
Facility-Wide Energy Forecaster
6-layer dilated TCN with 22 input channels combining building telemetry, weather forecast data, occupancy schedules, and historical consumption patterns. Output is a 24-point energy demand forecast (one per hour) used for demand-limiting strategies, pre-cooling/pre-heating scheduling, and battery storage dispatch where available.
Compilation constraints
All AxonML models targeting Hailo silicon are compiled under the fixed-function dataflow constraints: no dynamic control flow, no variable-length dimensions, all activations representable in INT8 after calibration, and no operations requiring dedicated softmax hardware (replaced with ReLU gating or depthwise convolution equivalents where necessary).
Silicon performance
Measured on production hardware via hailortcli benchmark with 5-second sustained inference. Device: Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289).
| Metric | Measured Value |
|---|---|
| FPS (hw_only) | 7,026.08 |
| FPS (streaming) | 6,996.48 |
| HW Latency | 0.178000 ms |
| Power (streaming avg) | 1.22500 W |
| Power (streaming max) | 1.22800 W |
| Power (idle) | 0.74979 W |
| Quantization | INT8 (post-training, DFC calibration) |
| DFC Compiler | 3.33.1 |
| HailoRT | 4.20.0 |
| Measured On | Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289) |
Deployment
Deployed as a single HEF binary. No ONNX runtime, TensorFlow Lite, or Python inference stack required at the edge.
| Target silicon | Hailo-8 |
| Measured on | Hailo-8 M.2 (P/N: HM218B1C2FAE, S/N: HLDDM2A234600289) |
| DFC compiler | 3.33.1 |
| Quantization | INT8 (post-training, production telemetry calibration) |
| Runtime | HailoRT (vendor runtime) |
| Edge platform | Raspberry Pi 5 + Hailo AI HAT+ (M.2 Key M) |
Deployment procedure
Copy the .hef binary to the target device. hailortcli run loads the HEF directly into the Hailo-8 dataflow engine over PCIe. Inference begins immediately with deterministic per-frame latency. No model conversion, graph optimization, or warmup phase required.
References
- Jewell, A. (2026). AxonML: A Pure-Rust Deep Learning Framework for Edge Inference. AutomataNexus LLC. Technical whitepaper.
- Hailo Technologies Ltd. (2024). Hailo Dataflow Compiler User Guide. DFC v3.33.1.
- Hailo Technologies Ltd. (2024). Hailo-8 Product Datasheet.
AutomataNexus LLC · Fort Wayne, Indiana · andrew.jewellsr@automatanexus.com
Andrew Jewell Sr. · ORCID 0009-0005-2158-7060
May 2026 · All rights reserved.