FLASH: Low-Latency Serverless Model Inference with Multi-Core Parallelism in Edge | IEEE Conference Publication | IEEE Xplore