biondizzle 0c02d84514 Add NaN/Inf detection in DeepseekV4Model.forward layer loop ...

- Checks every layer during prefill (not during cudagraph capture)
- is_current_stream_capturing() gate prevents CPU-GPU syncs during capture
- Prints amax every 10 layers for magnitude tracking
- Breaks on first NaN/Inf to avoid wasting compute

2026-05-17 23:37:12 +00:00

..

patches

Add NaN/Inf detection in DeepseekV4Model.forward layer loop

2026-05-17 23:37:12 +00:00

nvfp4_cutedsl.py

HOTFIX: remove NaN checks from run() — torch.isnan().any() does CPU-GPU sync, breaks cudagraph

2026-05-17 22:28:32 +00:00