Skip to content

Removed outdated warning about batch contamination #4423

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 6 commits into from
Nov 4, 2025
Merged

Removed outdated warning about batch contamination #4423

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
512769a
removed outdated warning
Harras3 Nov 3, 2025
6f2e1b0
Merge branch 'main' into Issue_4381
Harras3 Nov 3, 2025
aa7a7db
Merge branch 'main' into Issue_4381
Harras3 Nov 3, 2025
fb75c1b
Merge branch 'main' into Issue_4381
Harras3 Nov 3, 2025
98c435a
Merge branch 'main' into Issue_4381
Harras3 Nov 3, 2025
d372adc
Merge branch 'main' into Issue_4381
Harras3 Nov 4, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 0 additions & 3 deletions docs/source/reducing_memory_usage.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -90,9 +90,6 @@ from trl import SFTConfig
training_args = SFTConfig(..., packing=True, max_length=512)
```

> [!WARNING]
> Packing may cause batch contamination, where adjacent sequences influence one another. This can be problematic for some applications. For more details, see [#1230](https://github.com/huggingface/trl/issues/1230).

## Liger for reducing peak memory usage

> [Liger Kernel](https://github.com/linkedin/Liger-Kernel) is a collection of Triton kernels designed specifically for LLM training. It can effectively increase multi-GPU training throughput by 20% and reduce memory usage by 60%.
Expand Down