Expand description
LoRA (Low-Rank Adaptation) implementations for SONA
Two-tier LoRA system:
- MicroLoRA: Rank 1-2, per-request adaptation (<100μs)
- BaseLoRA: Rank 4-16, background adaptation (hourly)
Structs§
- Base
LoRA - Base LoRA for background adaptation
- LoRA
Engine - Combined LoRA engine managing both tiers
- LoRA
Layer - Single LoRA layer
- Micro
LoRA - Micro-LoRA for per-request adaptation
Constants§
- OPTIMAL_
BATCH_ SIZE - Optimal batch size for processing (benchmark-validated)