sm_120 SAME-L: AOT (graph-capturable) engines + tensorRT/README.md (#6) a05f2d4 cortexelus commited on 23 days ago
Rebuild medium DiT fp16-mixed engines (sm_90 + new sm_120): attention now fuses, 4.3x faster (#4) 06debc3 cortexelus commited on 24 days ago
