Hermes: Memory-Efficient Pipeline Inference for Large Models on Edge Devices

Devs

Hermes: Memory-Efficient Pipeline Inference for Large Models on Edge Devices | Read Paper on Bytez