Transferable Latency Prediction for Fast LLM Screening on Heterogeneous Edge Devices


This is a companion discussion topic for the original entry at https://arxiv.org/abs/2607.21602