Skip to content

docs: document training_regime field in the annotated model card #2656

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

Open
Sofeso wants to merge 1 commit into
base: main
Choose a base branch
from
Open

docs: document training_regime field in the annotated model card #2656

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
7 changes: 7 additions & 0 deletions docs/hub/model-card-annotated.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -174,6 +174,13 @@ _Write 1-2 sentences on what the training data is. Ideally this links to a Datas

_Detail tokenization, resizing/rewriting (depending on the modality), etc._

### Training Hyperparameters


`training_regime`

_The training regime, i.e. the numerical precision the model was actually trained with: `fp32`, `fp16` mixed precision, `bf16` mixed precision, `bf16` non-mixed precision, `fp16` non-mixed precision, or `fp8` mixed precision. This affects training speed and memory footprint, and in some cases the numerical stability of results, so it's useful for anyone trying to reproduce training or understand its resource requirements._

@meg-huggingface meg-huggingface Jul 21, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Since training regime is more broad, would suggest either

`numerical_precision`

_Provide the numerical precision the model was trained with (e.g., `fp32`, `fp16` mixed precision, `bf16` mixed precision, `bf16` non-mixed precision, `fp16` non-mixed precision, or `fp8` mixed precision). This affects training speed and memory footprint, and in some cases the numerical stability of results, so it's useful for anyone trying to reproduce training or understand its resource requirements._

Or go more broad:

`training_regime`

_Provide the training details that are useful for anyone trying to reproduce training or understand its resource requirements. This can include learning rate, batch size, and regularization method, among other hyperparameters. Numerical precision is particularly relevant here (e.g., `fp32`, `fp16` mixed precision, `bf16` mixed precision, `bf16` non-mixed precision, `fp16` non-mixed precision, or `fp8` mixed precision), as it affects training speed, memory footprint, and in some cases numerical stability._


### Speeds, Sizes, Times


Expand Down