Incredible stuff. Now I need a Navi fairy companion version.
Depth-aware light injection in TypeGPU I got a 448x448 monocular depth model down to ~8 ms on my M4 Pro across ~250 dispatches, which is fast enough to use in r...
fofr 称赞 TypeGPU 的深度感知光注入演示。该方案在 M4 Pro 上以约 8 ms 运行 448x448 单目深度模型,约 250 次调度,推理、光照与绘制共用同一命令编码器,无需 GPU 间同步。
fofr 称赞 TypeGPU 的深度感知光注入演示。该方案在 M4 Pro 上以约 8 ms 运行 448x448 单目深度模型,约 250 次调度,推理、光照与绘制共用同一命令编码器,无需 GPU 间同步。
Incredible stuff. Now I need a Navi fairy companion version.
来源:fofr· x.com