Skip to content

GitLab

Unverified Commit da885ed5 authored by Camille Zhong's avatar Camille Zhong Committed by GitHub
Browse files

fix tensor data update for gemini loss caluculation (#5442)

parent 8020f426
Show whitespace changes
Inline Side-by-side
Showing with 1 addition and 0 deletions
applications/Colossal-LLaMA-2/train.py
View file @ da885ed5
...@@ -56,6 +56,7 @@ def format_numel_str(numel: int) -> str: ...@@ -56,6 +56,7 @@ def format_numel_str(numel: int) -> str:
def all_reduce_mean(tensor: torch.Tensor) -> torch.Tensor: def all_reduce_mean(tensor: torch.Tensor) -> torch.Tensor:
dist.all_reduce(tensor=tensor, op=dist.ReduceOp.SUM) dist.all_reduce(tensor=tensor, op=dist.ReduceOp.SUM)
tensor = tensor.data
tensor.div_(dist.get_world_size()) tensor.div_(dist.get_world_size())
return tensor return tensor
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment