Skip to content
CUDA Lane

docs(fp8): correct TL;DR memory-math overclaim — bs4-fp8 ≈ 2× bs1-fp1… #154

Sign in to view logs

docs(fp8): correct TL;DR memory-math overclaim — bs4-fp8 ≈ 2× bs1-fp1…

docs(fp8): correct TL;DR memory-math overclaim — bs4-fp8 ≈ 2× bs1-fp1… #154

Run time

Learn about OS pricing on GitHub Actions
Job Run time
Run CUDA-only tests
1d 0h 0m 1s
1d 0h 0m 1s