Edge Deployment
Cross-source consensus on Edge Deployment from 1 sources and 5 claims.
1 sources · 5 claims
How it works
Benefits
Evidence quality
Highlighted claims
- The monitoring deployment uses an INT8 encoder, INT4 state head, and INT8 agitation MLP. — Mixed-Precision Information Bottlenecks for On-Device Trait-State Disentanglement in Bipolar Agitation Detection
- The continuous-monitoring footprint is reported as 617.1 KB. — Mixed-Precision Information Bottlenecks for On-Device Trait-State Disentanglement in Bipolar Agitation Detection
- Edge benchmarks were measured on Raspberry Pi Zero 2W with TensorFlow Lite and custom ARMv8-A NEON INT4 kernels. — Mixed-Precision Information Bottlenecks for On-Device Trait-State Disentanglement in Bipolar Agitation Detection
- Cortex-M7 deployment was not physically validated. — Mixed-Precision Information Bottlenecks for On-Device Trait-State Disentanglement in Bipolar Agitation Detection
- The MP-IB monitoring configuration is reported to run end-to-end in 23.4 ms on Raspberry Pi Zero 2W. — Mixed-Precision Information Bottlenecks for On-Device Trait-State Disentanglement in Bipolar Agitation Detection