Don't have Telegram yet? Try it now!
https://hgpu.org/?p=18979
A Unified Optimization Approach for CNN Model Inference on Integrated GPUs