Roast topics
Find topics
Roast it!
Roast topics
Find topics
Find it!
Login
From:
pytorch.org
(Uncensored)
subscribe
Accelerating 2K scale pre-training up to 1.28x with TorchAO, MXFP8 and TorchTitan on Crusoe B200 Cluster – PyTorch
https://pytorch.org/blog/accelerating-2k-scale-pre-training-up-to-1-28x-with-torchao-mxfp8-and-torchtitan-on-crusoe-b200-cluster/
links
backlinks
Tagged with:
blog
tldr: 1.22x – 1.28x training acceleration with MXFP8, equivalent convergence compared to BF16.
