Yard
heavy · gpu · 2026
LoRA-JIT
Page LoRA weights on a normal GPU so local models do not eat the machine.
Adapter routing, load what you need when you need it, and numbers that say if it is actually faster. Still in the lab.
- Year
- 2026
- Status
- lab
- Python
- PEFT
- FastAPI
- local LLM
Nearby on the board